Skip to content

MCP servers › io.github.hermoso-ai › hermoso-ai-ad-generator

MCP server · registry snapshot 2026-11-07 · last tested 2026-11-08

hermoso-ai-ad-generator MCP server

Registry name io.github.hermoso-ai/hermoso-ai-ad-generator

by hermoso-ai · GitHub account (the registry's namespace rule)repository: same ownerSource ↗Website ↗

AI ad generator: make image and video ads for your brand and research the ads competitors run. — the publisher's description

5 of 5 record checks metRead-only · tools listed 2026-11-08version 1.0.0JavaScriptstreamable HTTP188 tools

Summary

io.github.hermoso-ai/hermoso-ai-ad-generator is a remote MCP server listed in the official MCP registry at version 1.0.0, under the namespace of the GitHub account hermoso-ai. Its linked repository, github.com/hermoso-ai/hermoso, belongs to the same GitHub account. On 2026-11-08 ProtocolProbe connected to it in read-only mode and read its inventory: 188 tools, 6 resources and 0 prompts.

Record checks

5 of 5 record checks met · 4 do not apply

  • met: Registry entry is active and says how to run it
  • met: Source repository link opens
  • met: Repository is not archived
  • met: Code pushed to the repository in the 90 days before our check
  • met: Repository publishes a security policy (SECURITY.md)
  • not applicable: Package names its licence with a standard (SPDX) identifier
  • not applicable: The version the registry lists can be installed and is not withdrawn
  • not applicable: npm package carries build provenance (attestation)
  • not applicable: No known critical or high advisory in the package or its dependencies
  • Not ranked by downloads: a remote server publishes no download figure
  • Not ranked by stars: 6 stars (ranked from 10)

From public records, not a safety rating. Each check with its reason · Is it safe to install?

Operate this server? Correct a fact · README badge · Run the checks yourself · Claim it

Key facts

Runs
remote (hosted by the publisher)
Transport
streamable HTTP
Your own values
none declared (not a guarantee)
Sign-in
none needed to open a session
Repository
public
Tools
188
Last tested
2026-11-08

Tool list size

128,521 tokens on Claude Opus 5.5

Added to every request that loads this server's tools (188 tools). Counted with Anthropic's token counter on 2026-11-09, from the tool list we read on 2026-11-08. Median across 28,560 counted servers: 2,657 tokens. How it is counted

How a client gets in

Connects without signing in

Read from the server's public metadata on 2026-11-08 (app.hermoso.ai), in our read-only connection; no tool was called. It may have changed since. How it is read

Tool list at the three APIs

Not sent yet for this version (1.0.0).

Add to your client

Observed fact

This server accepted a session without any credential and lists 64 tools whose name or declaration suggests they change data (call_tool, set_auto_reload, create_brand, delete_brand, upload_file and 59 more). That is our reading of the names; we did not call them. Source: our connection, 2026-11-08.

How fresh this is

Live-tested
yes, a real session (not a text-only review)
Last tested
2026-11-08
Mode
read-only (no tool calls)
Latest version
1.0.0, published 2026-11-06 (registry, as of the 2026-11-07 snapshot)
First listed version
2026-11-06 (earliest version still in the registry; older removed versions would be earlier)
Registry data
snapshot 2026-11-07
Package and repository
checked 2026-11-07
Engine
mcp-runtime-engine@0.1.0 · sdk@1.32.0

Dead or stale signals

None found in the records we hold

Looked for and not found: deprecated in the registry; repository link does not open; repository archived; no push in over 365 days. Does not apply: listed version cannot be installed or is withdrawn. Each line is a dated fact from a public record. It does not say whether the server still works, and a record can change after the date shown.

Changes since the last refresh

New in the registry: not in the list of 2026-11-06, listed on 2026-11-07 with version 1.0.0 (published 2026-11-06).

Source: the official MCP Registry list on 2026-11-06 and on 2026-11-07.

What it is

What does this server offer?

Observed inventory
188tools
6resources
0prompts

None of its 188 tools are described as changing anything. For 188 the reading was uncertain, so no mark is shown.

  • enable_toolson 4 other servers

    “LIST a group of tools that is not in this session's roster.”

    Publisher declares: read-only no · destructive no · idempotent yes · open-world no

    1 input, 1 required
    • groupsarrayrequired“Groups to switch on, e.g.”
  • “Search EVERY Hermoso tool — your starting list is deliberately short, and everything else in the product is here — by name, task or group.”

    Publisher declares: read-only yes · destructive no · idempotent yes · open-world no

    4 inputs
    • querystring“words from the task or the tool name, e.g.”
    • groupstring“limit to one group: core, research, create, channels, channel_admin, analytics, ads, files, workspace”
    • limitnumber“how many to return (default 12, max 40)”
    • onlyHealthyboolean“leave out tools that are failing their recent calls or that need a connector this workspace has not made.”
  • “Run ANY Hermoso tool by name — including the paid-campaign, analytics and channel-admin tools that are not in this session's starting list — with the same permissions, the same account and the same r…”

    Publisher declares: read-only no · destructive yes · idempotent no · open-world yes

    2 inputs, 1 required
    • namestringrequired“name”
    • argsobject“the tool's arguments as an object, exactly as its own schema takes them”
  • hermoso_capabilitieson 3 other servers

    “Probe what this Hermoso account can do RIGHT NOW: available image/video model ids + their credit costs (estimates: a render is charged its actual cost), aspect ratios, video durations, the recipe ids…”

    Publisher declares: read-only yes · destructive no · idempotent yes · open-world no

    Takes no inputs.

  • hermoso_creditson 3 other servers

    “Return the credit balance, recent priced calls with their credits, and costModel (the one-sentence rule of what costs credits).”

    Publisher declares: read-only yes · destructive no · idempotent yes · open-world no

    Takes no inputs.

  • “Out of credits?”

    Publisher declares: read-only no · destructive no · idempotent no · open-world yes

    3 inputs
    • packstring“the pack id to buy (e.g.”
    • confirmboolean“set true to actually charge the saved card for `pack` (required for the one-click charge; ignored on the checkout-link path)”
    • quote_tokenstring“the quoteToken returned by the quote step — REQUIRED (with confirm:true) to charge; it binds the exact pack + price you quoted (10-minute validity) and makes a…”
  • “Report a bug in Hermoso to the team.”

    Publisher declares: read-only no · destructive no · idempotent no · open-world yes

    3 inputs, 2 required
    • summarystringrequired“one-line summary of the bug”
    • detailsstringrequired“what you were doing, the tool + arguments you called, what you expected, and what actually happened (paste the exact error)”
    • severitystring“high = blocks the task or loses paid work; medium = wrong output but workable; low = cosmetic”
  • request_featureon 14 other servers

    “Ask the Hermoso team for a capability that doesn't exist yet.”

    Publisher declares: read-only no · destructive no · idempotent no · open-world yes

    2 inputs, 2 required
    • summarystringrequired“one line: the capability you need”
    • detailsstringrequired“what the user was actually trying to achieve, why the current tools couldn't do it, and what you'd expect the capability to do”
  • billing_statuson 9 other servers

    “Show this account's billing at a glance: current plan (quote plan.priceUsd per plan.period), credit balance, whether auto-reload is on, whether a card is on file, and whether YOU (this key) have ADMI…”

    Publisher declares: read-only yes · destructive no · idempotent yes · open-world no

    Takes no inputs.

  • “Change the SUBSCRIPTION plan (admin only).”

    Publisher declares: read-only no · destructive no · idempotent no · open-world yes

    4 inputs
    • planstring“the plan id to move to (e.g.”
    • periodstring“monthly only ('yr' buys monthly)”
    • confirmboolean“true (with quote_token) to make a quoted upgrade”
    • quote_tokenstring“the quoteToken from the quote step”
  • set_auto_reloadon 3 other servers

    “Turn automatic credit reloads on or off (admin only): when the balance drops below a threshold, the card on file is charged for a top-up pack — SERVER-SIDE, even with no app open.”

    Publisher declares: read-only no · destructive no · idempotent yes · open-world no

    3 inputs, 1 required
    • enabledbooleanrequired“true to turn auto-reload on, false to turn it off”
    • thresholdCreditsinteger“reload when the balance drops below this many credits”
    • reloadCreditsinteger“how many credits to add each reload — must match a credit pack size (see buy_credits)”
  • “List every profile on this account (id + name; a profile is a brand, a creator or a personal workspace) and which one this connection acts on, PLUS any profile another account shared with you.”

    Publisher declares: read-only yes · destructive no · idempotent yes · open-world no

    Takes no inputs.

Show 88 more tools
  • “Pin which profile this connection acts on.”

    Publisher declares: read-only no · destructive no · idempotent yes · open-world no

    1 input, 1 required
    • brandstringrequired“profile id (e.g.”
  • “Add a NEW profile (a brand, client, creator or personal workspace) and switch to it.”

    Publisher declares: read-only no · destructive no · idempotent yes · open-world no

    2 inputs, 1 required
    • namestringrequired“the name for the new profile (a brand, client or creator name)”
    • activateboolean“switch this connection to the new profile (default true); everything you do next scopes to it”
  • delete_brandon 4 other servers

    “PERMANENTLY delete a profile and EVERYTHING in it — brand details, memory, swipefile, Library creations, generated assets, avatars, skills, playbooks, chats — and disconnect its connected accounts.”

    Publisher declares: read-only no · destructive yes · idempotent yes · open-world yes

    4 inputs, 1 required
    • brandstringrequired“profile id or exact name from list_brands”
    • confirmboolean“REQUIRED true — this destroys the whole profile and cannot be undone”
    • confirmNamestring“the workspace's EXACT name, required when it is not empty — copy it from the inventory this tool returned, after the user has agreed to it”
    • confirmConnectorsnumber“the number of connected accounts the inventory reported, required when there is at least one — the user must specifically agree to losing them, because reconne…”
  • “Persist an ARBITRARY user file (image, video, audio, PDF or document, up to 150MB) into Hermoso and get back a durable public URL that EVERY publish, schedule and ad-build tool accepts.”

    Publisher declares: read-only no · destructive no · idempotent no · open-world yes

    4 inputs
    • getUploadUrlboolean“ASK FOR A ONE-TIME UPLOAD URL instead of uploading now — use this whenever the file is on the user’s machine and you can run a shell or an HTTP request.”
    • urlstring“a PUBLIC http(s) URL Hermoso fetches server-side (private/internal addresses are refused, and every redirect hop is re-checked).”
    • dataUristring“base64 data: URI of the file bytes (data:<mime>;base64,<…>) — bytes travel over this connection, so keep it small”
    • namestring“original file name — helps pick the right extension”
  • save_to_driveon 3 other servers

    “Save a Hermoso render — or ANY file — into the user’s connected Google Drive.”

    Publisher declares: read-only no · destructive no · idempotent no · open-world yes

    4 inputs
    • urlstring“a single Hermoso render URL to save”
    • urlsarray“several render URLs (up to 20) to save in one call”
    • folderstring“Drive folder name to save into (created if new)”
    • namestring“file name (single save)”
  • import_from_cloudon 3 other servers

    “Pull the files in a Google Drive or OneDrive FOLDER into this profile's Library, so they can be used like anything rendered here — published, scheduled, cloned, used as a product photo or a reference.”

    Publisher declares: read-only no · destructive no · idempotent no · open-world yes

    3 inputs, 1 required
    • providerstringrequired“which cloud — `drive` is Google Drive, `onedrive` is Microsoft OneDrive”
    • folderIdstring“the folder to import, from list_drive_files / list_onedrive_files (onlyFolders:true).”
    • limitnumber“how many files to bring across in this call (default 10, max 25).”
  • list_drive_fileson 3 other servers

    “List the Google Drive files & folders Hermoso can reach: the ones it created plus any the user picked with the Google file picker in the app (drive.file, never their whole Drive).”

    Publisher declares: read-only yes · destructive no · idempotent yes · open-world yes

    7 inputs
    • querystring“only files whose name contains this”
    • folderIdstring“list the contents of this folder id”
    • recursiveboolean“with folderId: the whole tree”
    • onlyFoldersboolean“list folders only”
    • pageSizenumber“rows per page (1–200, default 50)”
    • pageTokenstring“cursor from a previous call”
    • includeTrashedboolean“include trashed files (default false)”
  • get_drive_fileon 3 other servers

    “Fetch one Drive file’s metadata — name, type, size, modified time, a webViewLink to open it and a webContentLink to download it.”

    Publisher declares: read-only yes · destructive no · idempotent yes · open-world yes

    1 input, 1 required
    • fileIdstringrequired“the Drive file id (from list_drive_files)”
  • update_drive_fileon 4 other servers

    “Update a Drive file: rename (name), move it into a folder (moveToFolderId, optionally removeFromFolderId to move OUT of the old one), or trash / untrash it (trash:true|false).”

    Publisher declares: read-only no · destructive yes · idempotent yes · open-world yes

    5 inputs, 1 required
    • fileIdstringrequired“the Drive file id”
    • namestring“new name”
    • moveToFolderIdstring“folder id to move the file into (from create_drive_folder / list_drive_files)”
    • removeFromFolderIdstring“the old parent folder id to remove (when moving)”
    • trashboolean“true -> move to Trash; false -> restore from Trash”
  • delete_drive_fileon 4 other servers

    “Delete a Drive file.”

    Publisher declares: read-only no · destructive yes · idempotent yes · open-world yes

    3 inputs, 1 required
    • fileIdstringrequired“the Drive file id”
    • permanentboolean“true = delete forever; default trashes (recoverable)”
    • confirmboolean“REQUIRED true”
  • create_drive_folderon 4 other servers

    “Create a folder in the user’s Google Drive (optionally nested under parentId) to organize saved files.”

    Publisher declares: read-only no · destructive no · idempotent no · open-world yes

    2 inputs, 1 required
    • namestringrequired“folder name”
    • parentIdstring“parent folder id for a nested folder (default: Drive root)”
  • “Create a new Google Spreadsheet in the user’s Drive and optionally fill it with rows — e.g.”

    Publisher declares: read-only no · destructive no · idempotent no · open-world yes

    2 inputs
    • titlestring“spreadsheet title”
    • rowsarray“rows to write — array of row arrays; first row = headers”
  • append_to_sheeton 3 other servers

    “Append rows to a Google Sheet Hermoso can reach — one it created (pass the spreadsheetId from create_sheet) or one the user handed over with the Google file picker in the app (find its id with list_d…”

    Publisher declares: read-only no · destructive no · idempotent no · open-world yes

    3 inputs, 2 required
    • spreadsheetIdstringrequired“the spreadsheet id from create_sheet”
    • rowsarrayrequired“rows to append — array of row arrays”
    • rangestring“range to append at (default A1 / first sheet)”
  • “Read cells from a Google Sheet Hermoso can reach — one it created, or one the user handed over with the Google file picker in the app (that is how an EXISTING spreadsheet becomes readable; find its i…”

    Publisher declares: read-only yes · destructive no · idempotent yes · open-world yes

    3 inputs
    • spreadsheetIdstring“the spreadsheet id (from create_sheet)”
    • sheetUrlstring“a Google Sheets URL to read — the spreadsheet id is extracted from it”
    • rangestring“A1 range, e.g.”
  • export_swipefile_deckon 3 other servers

    “Turn a SWIPEFILE COLLECTION into a real Google Slides deck — one slide per saved ad, carrying the creative, the brand, the ad copy, the run dates with the run length, and the platform.”

    Publisher declares: read-only no · destructive no · idempotent no · open-world yes

    3 inputs
    • collectionstring“the swipefile collection to export, by name or id (default: the first collection)”
    • titlestring“deck title (default: the collection name)”
    • limitnumber“max ads to include, 1-60 (default 30)”
  • list_notion_pageson 3 other servers

    “Find the Notion pages and databases the user shared with Hermoso, by title, newest edited first.”

    Publisher declares: read-only yes · destructive no · idempotent yes · open-world yes

    4 inputs
    • querystring“words in the title”
    • typestring“limit results to pages or to databases; omit for both”
    • limitnumber“1-100, default 25”
    • cursorstring“from the previous page”
  • save_to_notionon 3 other servers

    “Save to the user's Notion as a NEW page: a research report, competitor teardown or campaign report as `content` (markdown: headings, lists, links, and an image or video URL on its own line becomes th…”

    Publisher declares: read-only no · destructive no · idempotent no · open-world yes

    8 inputs
    • titlestring“title of the new Notion page”
    • contentstring“markdown body”
    • swipefileCollectionstring“swipefile collection name or id”
    • includeScheduleboolean“add the scheduled posts”
    • parentPageIdstring“page id or link to save under”
    • databaseIdstring“database id or link: saves as a new row”
    • dataSourceIdstring“data source id to save into as a new row; pass only one of parentPageId, databaseId, dataSourceId”
    • iconstring“one emoji”
  • append_to_notion_pageon 3 other servers

    “Add to the END of an existing Notion page (pageId or its link, from list_notion_pages): markdown content, a swipefileCollection, or includeSchedule:true.”

    Publisher declares: read-only no · destructive no · idempotent no · open-world yes

    4 inputs, 1 required
    • pageIdstringrequired“page id or link”
    • contentstring“markdown to add”
    • swipefileCollectionstring“swipefile collection name or id to append, with each ad's creative, copy, dates and link”
    • includeScheduleboolean“true appends the scheduled posts too”
  • read_notion_pageon 3 other servers

    “Read a Notion page the user shared with Hermoso, as markdown (pageId or a Notion link).”

    Publisher declares: read-only yes · destructive no · idempotent yes · open-world yes

    1 input, 1 required
    • pageIdstringrequired“page id or Notion link”
  • export_schedule_to_notionon 3 other servers

    “Export the posting schedule to a Notion database, one row per post: date, channels, caption, media link and status.”

    Publisher declares: read-only no · destructive no · idempotent no · open-world yes

    5 inputs
    • parentPageIdstring“Notion page to create the new database under; omit for a private page in the workspace”
    • databaseIdstring“an existing database id or link”
    • dataSourceIdstring“data source id of an existing database to write rows into, matched to columns by name”
    • titlestring“name for a new database”
    • historyDaysnumber“0-90 days of published posts to include”
  • query_notion_databaseon 3 other servers

    “Read the ROWS of a Notion database the user shared with Hermoso — e.g.”

    Publisher declares: read-only yes · destructive no · idempotent yes · open-world yes

    9 inputs
    • databaseIdstring“database id or link (list_notion_pages type:"database")”
    • dataSourceIdstring“data source id from an earlier read, as an alternative to databaseId”
    • matchobject“{"Column": value} — equality for status/select/checkbox/number/date, contains for text/multi-select”
    • filterobject“a raw Notion filter object”
    • sortBystring“column name to sort by, or created_time or last_edited_time”
    • directionstring“ascending (default) or descending”
    • limitnumber“1-100, default 50”
    • cursorstring“nextCursor from the previous page of rows”
    • archivedboolean“read archived rows instead”
  • update_notion_rowon 3 other servers

    “Change properties on ONE row of a Notion database (pageId from query_notion_database) — e.g.”

    Publisher declares: read-only no · destructive yes · idempotent yes · open-world yes

    4 inputs, 1 required
    • pageIdstringrequired“the row id or link”
    • propertiesobject“{"Column": value}”
    • trashboolean“true moves the row to Notion trash (needs confirm); false restores it”
    • confirmboolean“true confirms moving the row to trash; not needed for other changes”
  • “Create a new Google Doc in the user’s Drive with a title + optional body text — e.g.”

    Publisher declares: read-only no · destructive no · idempotent no · open-world yes

    2 inputs
    • titlestring“document title”
    • textstring“body text to insert”
  • append_to_docon 3 other servers

    “Append text to the end of a Google Doc Hermoso can reach — one it created (pass the documentId from create_doc) or one the user handed over with the Google file picker in the app (find its id with li…”

    Publisher declares: read-only no · destructive no · idempotent no · open-world yes

    3 inputs, 1 required
    • documentIdstringrequired“the document id from create_doc”
    • textstring“text to append at the end of the doc (optional when a dropdown is passed)”
    • dropdownobject“a dropdown chip to append after the text”
  • “Read the text of a Google Doc Hermoso can reach — one it created, or one the user handed over with the Google file picker in the app (that is how an EXISTING doc becomes readable; find its id with li…”

    Publisher declares: read-only yes · destructive no · idempotent yes · open-world yes

    2 inputs
    • documentIdstring“the document id (from create_doc)”
    • docUrlstring“a Google Docs URL to read — the document id is extracted from it”
  • list_sheet_tabson 3 other servers

    “The tabs in a Google Spreadsheet, each with its name, numeric sheetId, row/column count and position.”

    Publisher declares: read-only yes · destructive no · idempotent yes · open-world yes

    2 inputs
    • spreadsheetIdstring“the spreadsheet id (from create_sheet, or list_drive_files for one the user picked)”
    • sheetUrlstring“a Google Sheets URL — the id is extracted from it”
  • update_sheeton 4 other servers

    “CORRECT cells in a Google Sheet — write values to an exact range, overwriting whatever is there.”

    Publisher declares: read-only no · destructive yes · idempotent yes · open-world yes

    7 inputs
    • spreadsheetIdstring“Google spreadsheet id; use this or sheetUrl”
    • sheetUrlstring“Google Sheets link, instead of spreadsheetId”
    • rangestring“A1 range or anchor cell, e.g.”
    • valuesarray“array of row arrays to write”
    • updatesarray“write SEVERAL disjoint ranges in one call, instead of range+values”
    • valueInputOptionstring“USER_ENTERED (default) parses formulas, dates and numbers the way typing them would; RAW stores every value as literal text”
    • confirmboolean“required only when the target range already holds values”
  • clear_sheet_rangeon 3 other servers

    “Empty a range of cells in a Google Sheet, leaving the rows themselves in place.”

    Publisher declares: read-only no · destructive yes · idempotent yes · open-world yes

    5 inputs, 1 required
    • spreadsheetIdstring“Google spreadsheet id; use this or sheetUrl”
    • sheetUrlstring“Google Sheets link, instead of spreadsheetId”
    • rangestringrequired“the range to clear, e.g.”
    • confirmboolean“true clears the range; call without it first to get the filled-cell count”
    • confirmCellsnumber“echo back the filled-cell count the unconfirmed call reported”
  • manage_sheet_tabson 3 other servers

    “Add, rename or delete a tab in a Google Spreadsheet.”

    Publisher declares: read-only no · destructive yes · idempotent no · open-world yes

    8 inputs, 1 required
    • spreadsheetIdstring“Google spreadsheet id; use this or sheetUrl”
    • sheetUrlstring“Google Sheets link, instead of spreadsheetId”
    • actionstringrequired“add, rename or delete a tab”
    • tabstring“which tab — its title or numeric sheetId (rename / delete)”
    • titlestring“the name for the new tab (action:"add")”
    • newTitlestring“what to rename the tab to (action:"rename")”
    • confirmboolean“true deletes a tab that holds data; needs confirmCells”
    • confirmCellsnumber“the filled-cell count an unconfirmed delete reported”
  • format_sheeton 4 other servers

    “Make an exported sheet readable: bold the header row, FREEZE it so it stays visible while scrolling, and auto-size the columns so nothing is cut off.”

    Publisher declares: read-only no · destructive yes · idempotent yes · open-world yes

    6 inputs
    • spreadsheetIdstring“Google spreadsheet id; use this or sheetUrl”
    • sheetUrlstring“Google Sheets link, instead of spreadsheetId”
    • tabstring“tab title or numeric sheetId (default: the first tab)”
    • boldHeaderboolean“bold the header row; default true, false skips it”
    • freezeRowsnumber“how many top rows to freeze (default 1, 0 = none)”
    • autoResizeboolean“auto-size columns to fit their content; default true, false skips it”
  • “EDIT a Google Doc — the correction append_to_doc cannot make.”

    Publisher declares: read-only no · destructive yes · idempotent yes · open-world yes

    7 inputs
    • documentIdstring“the document id (from create_doc, or list_drive_files for one the user picked)”
    • docUrlstring“a Google Docs URL — the id is extracted from it”
    • replacementsarray“find/replace pairs, applied in order”
    • rewritestring“replace the WHOLE body with this text ("" empties the doc)”
    • confirmboolean“true runs a whole-body rewrite after the character count is shown”
    • confirmCellsnumber“echo back the character count the unconfirmed call reported (rewrite only)”
    • dropdownsarray“dropdown chips to set, or dropdowns to change”
  • save_to_onedriveon 3 other servers

    “Save a Hermoso render — or ANY file — into the user’s connected Microsoft OneDrive.”

    Publisher declares: read-only no · destructive no · idempotent no · open-world yes

    4 inputs
    • urlstring“a single Hermoso render URL to save”
    • urlsarray“several render URLs (up to 20) to save in one call”
    • folderstring“OneDrive folder name to save into (created if new)”
    • namestring“file name (single save)”
  • list_onedrive_fileson 3 other servers

    “List files & folders in the user’s OneDrive — the root by default, a folder’s contents (folderId), or a name search (query).”

    Publisher declares: read-only yes · destructive no · idempotent yes · open-world yes

    5 inputs
    • querystring“search — only items whose name matches this”
    • folderIdstring“list the contents of this folder id”
    • onlyFoldersboolean“list folders only”
    • pageSizenumber“rows per page (1–200, default 50)”
    • pageTokenstring“cursor from a previous call”
  • get_onedrive_fileon 3 other servers

    “Fetch one OneDrive item’s metadata — name, type, size, modified time, a webViewLink to open it and a webContentLink to download it.”

    Publisher declares: read-only yes · destructive no · idempotent yes · open-world yes

    1 input, 1 required
    • fileIdstringrequired“the OneDrive item id (from list_onedrive_files)”
  • update_onedrive_fileon 3 other servers

    “Update a OneDrive item: rename (name) and/or move it into a folder (moveToFolderId).”

    Publisher declares: read-only no · destructive yes · idempotent yes · open-world yes

    3 inputs, 1 required
    • fileIdstringrequired“the OneDrive item id”
    • namestring“new name”
    • moveToFolderIdstring“folder id to move the item into (from create_onedrive_folder / list_onedrive_files)”
  • delete_onedrive_fileon 3 other servers

    “Delete a OneDrive item — it moves to the OneDrive recycle bin (recoverable there).”

    Publisher declares: read-only no · destructive yes · idempotent yes · open-world yes

    2 inputs, 1 required
    • fileIdstringrequired“the OneDrive item id”
    • confirmboolean“REQUIRED true”
  • share_onedrive_fileon 3 other servers

    “Create a SHARING LINK to a OneDrive file or folder (fileId from list_onedrive_files): type view (read-only, default), edit, or embed (an <iframe> link, personal OneDrive only); scope anonymous (anyon…”

    Publisher declares: read-only no · destructive yes · idempotent yes · open-world yes

    6 inputs, 1 required
    • fileIdstringrequired“OneDrive file or folder id from list_onedrive_files”
    • typestring“view (default), edit, or embed (personal OneDrive only)”
    • scopestring“anonymous (default, anyone with the link), organization (work or school), or users (invited people only)”
    • expiresAtstring“e.g.”
    • passwordstring“personal OneDrive with Microsoft 365”
    • confirmboolean“required for an anonymous edit link”
  • onedrive_file_permissionson 3 other servers

    “action list (default): every sharing link and person with access to a OneDrive file or folder: link URL, view or edit, scope, expiry, password, whether it is inherited from a parent folder.”

    Publisher declares: read-only no · destructive yes · idempotent yes · open-world yes

    4 inputs, 1 required
    • fileIdstringrequired“OneDrive file or folder id”
    • actionstring“list (default) or remove”
    • permissionIdstring“for remove”
    • confirmboolean“required true for remove”
  • convert_onedrive_fileon 3 other servers

    “Turn a file already in the user’s OneDrive into a PDF or a JPG — Microsoft does the conversion on its own servers, so nothing is re-encoded here and nothing is lost in a screenshot.”

    Publisher declares: read-only no · destructive no · idempotent no · open-world yes

    4 inputs, 1 required
    • fileIdstringrequired“the OneDrive item id, from list_onedrive_files”
    • formatstring“default pdf”
    • widthnumber“REQUIRED for jpg — output width in pixels”
    • heightnumber“REQUIRED for jpg — output height in pixels”
  • create_onedrive_folderon 3 other servers

    “Create a folder in the user’s OneDrive (optionally nested under parentId) to organize saved files.”

    Publisher declares: read-only no · destructive no · idempotent no · open-world yes

    2 inputs, 1 required
    • namestringrequired“folder name”
    • parentIdstring“parent folder id for a nested folder (default: OneDrive root)”
  • “Plan an ad: turns a brand and a product or brief into an ad concept, with copy variants (headline, primary text, call to action) and either an image prompt or a timed video storyboard, plus the recip…”

    Publisher declares: read-only no · destructive no · idempotent no · open-world yes

    13 inputs, 1 required
    • brandstring or object“brand name, or a brand profile object {name,domain,category,palette,products,…}.”
    • productstringrequired“what to advertise + any angle/offer the user specified”
    • formatstring“'image', 'video', or 'auto' when unspecified”
    • durationSecondsnumber“For video: the total length the user asked for, in seconds (30 for "a 30 second ad").”
    • hookstring“force the VISUAL scroll-stop mechanic the opening beat is built on — a hook id from list_hooks (e.g.”
    • settingstring“force the WHERE — a setting id from list_hooks (e.g.”
    • recipestring“a recipe id from hermoso_capabilities to force an archetype”
    • referencestring“a reference to clone: an ad-library link (Facebook Ad Library, LinkedIn Ad Library, Google Ads Transparency — its real copy/advertiser are fetched) OR a VIDEO …”
    • languagestring“output language for the ad copy (e.g.”
    • talentstring“who is on camera; omit to follow the format.”
    • conceptsarray“several variants: one line per concept (a different angle each); built exactly”
    • draftobject“Only after a video refusal that offered a lighter draft: the {model, durationSeconds} that refusal named.”
    • insertsarray“VIDEO ONLY: the user's own SCREENS cut in at these seconds (each item's look, kind, start and end say how), the voice running underneath.”
  • “Remake a video you like FOR THIS BRAND from its link — a TikTok, Instagram Reel, Facebook video or reel, X post, or a direct video file URL.”

    Publisher declares: read-only no · destructive no · idempotent no · open-world yes

    6 inputs, 1 required
    • urlstringrequired“the video to clone — a TikTok / Instagram Reel / Facebook / X link, or a direct https video file URL (not YouTube, which cannot be downloaded); a URL, not a lo…”
    • productstring“what the new ad sells, plus any angle or offer; omit to use the saved brand's product”
    • changesstring“what to change or keep from the original, in the user's words (e.g.”
    • brandstring or object“brand name or profile object; OMIT to use the workspace's saved brand (see get_brand)”
    • durationSecondsnumber“override the length in seconds; omit to match the original”
    • languagestring“language for the new ad's script and copy — default English”
  • “Paste a product page link (Shopify, Amazon or any store page) and get a creator-style video ad planned for that product.”

    Publisher declares: read-only no · destructive no · idempotent no · open-world yes

    5 inputs, 1 required
    • urlstringrequired“the product page link, https://…”
    • stylestring“ugc_selfie (talking to camera, default), ugc_unboxing, ugc_demo (how-to), ugc_tryon (try-on or haul), ugc_product (product close-ups, no face), ugc_shopping (i…”
    • durationSecondsnumber“length in seconds, 4-180; default 15”
    • saystring“what the creator should say, or the angle, in the user's words”
    • languagestring“language for the script and copy, default English”
  • generate_imageon 59 other servers

    “Use this when the user asks for an image ad, ad creative, product photo or social post graphic for their business or product, or an ANIMATED GIF (format gif) or SVG file (format svg).”

    Publisher declares: read-only no · destructive no · idempotent no · open-world yes

    15 inputs
    • promptstring“REQUIRED on every model EXCEPT the pose rows below.”
    • refImagesarray“local file paths or URLs of product/logo references to composite in.”
    • useBrandboolean“default true: with no refImages, the server hydrates the SAVED brand’s product/logo references so the output lands on-brand; pass false for a pure prompt-only …”
    • rawboolean“RAW MODEL ACCESS: run the caller’s prompt on the named model with no Hermoso adjustments at all — the prompt reaches the provider byte-identical (no hex-to-col…”
    • aspectRatiostring“e.g.”
    • modelstring“image model id from hermoso_capabilities.”
    • jobstring“what the picture is for.”
    • formatstring“the file to make (default png).”
    • framesinteger“format gif only: frames in the loop, 2 to 16 (default 9)”
    • frameMsinteger“format gif only: milliseconds per frame, 20 to 2000, rounded to 10 (default 200; 9 frames x 200 = a 1.8 s loop)”
    • complexitystring“gpt-image-2.6 / 2.6-flare only: how much the model plans before drawing (default high, best for dense text and complex layouts; low suits a small focused chang…”
    • imageSizestring“pixel-size preset for models that support it: 1K/2K, and 4K on the models hermoso_capabilities lists with a 4K imageSize price (a 4K ask on any other model is …”
    • fixLabelboolean“default true: when the saved brand's product photo rides in this render, the product's label on the finished image is READ and compared with the photo, and re-…”
    • logoPlacementstring“overlay: real logo file laid flat.”
    • maskstring“MASKED EDIT — change ONE region and keep the rest: a mask (path or URL, any size) for refImages[0].”
  • “EDIT an existing image in place with a plain-language instruction and keep everything else: 'make the headline bigger', 'add our logo bottom right', 'swap the background for a kitchen', 'remove the p…”

    Publisher declares: read-only no · destructive no · idempotent no · open-world yes

    10 inputs, 2 required
    • imagestringrequired“the image to edit: URL, Library item URL, upload_file URL or local path”
    • instructionstringrequired“the change to make, in plain words (pass the user’s own words for a removal or plain photo edit)”
    • removalboolean“true when the edit REMOVES text, branding, a logo, a watermark, a person or an object, so the brand name and logo are not re-added”
    • logoboolean
    • maskstring“optional mask image (URL or local path) marking the region to change: transparent = change, or white = change on an opaque mask”
    • markupstring“MARKUP-GUIDED EDIT: drawn annotations (circles, arrows, scribbles) showing what to change, as a URL or local path: a transparent PNG of just the marks, or a co…”
    • modelstring“the editor, only when the user names one: gpt-image-2.6, gpt-image-2.6-flare (its own faster editor) or gpt-image-2.5.”
    • complexitystring“gpt-image-2.6 editors only: how much it plans before drawing (default high; low suits a small focused change)”
    • dryRunboolean“true = return what this edit usually costs in credits and render nothing”
    • fixLabelboolean“false = leave the product label exactly as rendered: no product lookup, no label check, no re-print (default on when the ad shows the brand’s saved product)”
  • headline_variantson 1 other server

    “Turn ONE finished static ad into several copies that differ ONLY in the headline, for an A/B test: same picture, product, layout, colours and every other line.”

    Publisher declares: read-only no · destructive no · idempotent no · open-world yes

    5 inputs, 1 required
    • imagestringrequired“the finished static ad: URL, Library item URL, upload_file URL or local path”
    • headlinesarray“your own headlines to test (up to 10); omit to have them written”
    • countinteger“how many headlines to write when `headlines` is omitted (default 5)”
    • briefstring“what to test, e.g.”
    • fixLabelboolean“false = leave the product label exactly as rendered: no product lookup, no label check, no re-print (default on when the ad shows the brand’s saved product)”
  • “Re-lay out ONE finished static ad for other placements: the same ad, product, copy (word for word), logo and style, recomposed natively for each canvas rather than cropped.”

    Publisher declares: read-only no · destructive no · idempotent no · open-world yes

    3 inputs, 1 required
    • imagestringrequired“the finished static ad: URL, Library item URL, upload_file URL or local path”
    • aspectRatiosarray“target canvases (default 1:1, 4:5, 9:16)”
    • fixLabelboolean“false = leave the product label exactly as rendered: no product lookup, no label check, no re-print (default on when the ad shows the brand’s saved product)”
  • “Translate the on-image text of ONE finished static ad into other languages and keep everything else: same picture, layout, typeface, colours, logo and product.”

    Publisher declares: read-only no · destructive no · idempotent no · open-world yes

    3 inputs, 2 required
    • imagestringrequired“the finished static ad: URL, Library item URL, upload_file URL or local path”
    • languagesarrayrequired“target languages, by name”
    • fixLabelboolean“false = leave the product label exactly as rendered: no product lookup, no label check, no re-print (default on when the ad shows the brand’s saved product)”
  • make_thumbnailon 1 other server

    “Render a click-driving YOUTUBE / Shorts / Instagram THUMBNAIL or video cover through the full production pipeline (concept, casting, scene, render, tweaks, text), not a bare image prompt.”

    Publisher declares: read-only no · destructive no · idempotent no · open-world yes

    33 inputs
    • frameworkstring“concept framework id (default 'posed_portrait') — before_after · social_ui · three_step · screenshot · posed_portrait · posed_action · specific_day · graphical…”
    • frameworkRequestedboolean“true ONLY when the USER named this framework — it is what authorizes a text-carrying framework (social_ui / news_clip / specific_day / map_aerial) to bake its …”
    • sceneBriefstring“what the thumbnail depicts — the concept in one dense sentence, rendered exactly”
    • topicstring“the video's topic — used to pick the hero object when you don't name keyElements”
    • headlinestring“2–4 word headline.”
    • headlineLinesarray“explicit headline lines (up to 3) — overrides splitting `headline` on newlines”
    • bakeTextboolean“default false.”
    • bakedUiTextstring“short label for a text-carrying framework (a chat bubble, a DAY N badge, a news lower-third, a map callout) — needs frameworkRequested:true”
    • overlayStylestring“headline style: beast (default), fire, neon-lime, clean-glass, marker, or your own CSS declarations”
    • fontstring“headline font: Anton (default) or any Google Fonts family”
    • headlinePlacestring“bottom (default), top, center, or a 0-1 fraction from the top; never over the face”
    • faceImagesarray“up to 3 face photos (URLs or local paths) — each becomes a locked CHARACTER identity, in order”
    • peoplearray“people described in prose instead of by photo (each still gets the chosen expression)”
    • castGenericPersonboolean“pass true only after the user has explicitly chosen a generated stranger over their own face”
    • emotionstring“the expression on the face (default 'shock') — shock · hype · fear · confusion · determination · smug · charisma · disgust · awe · rage · laugh, or your own ph…”
    • emotionsarray“render one variant per emotion (variants = emotions × takes, max 16)”
    • takesnumber“camera takes per emotion, 1–4: designed framing / low-angle hero / extreme close-up / wide dutch tilt”
    • variantsnumber“how many thumbnails to render (default 1, max 16).”
    • aspectRatiostring“'16:9' (YouTube, default) / '9:16' (Shorts) / '4:5' (Instagram) / '4:3' / '1:1'”
    • keyElementsstring“signature props / effects that make it pop — oversized, flying toward camera”
    • locationstring“place, time of day, weather, atmosphere”
    • compositionstring“override the default large-foreground-subject composition”
    • backgroundstring“override the default bold saturated colour-field background”
    • rimColorstring“colored back+hair light — ONLY when the user names one: 'ice-blue' / 'neon-magenta' / 'toxic-lime' / 'amber-gold' / 'pure-white'”
    • restrainedGradeboolean“true for a calm / premium / muted look instead of the default punchy poster grade”

    and 8 more not listed here.

  • generate_voiceon 4 other servers

    “RAW speech as an MP3: one voice reading a script (text), or a DIALOGUE of two or more people (lines), each in their own voice.”

    Publisher declares: read-only no · destructive no · idempotent no · open-world yes

    5 inputs
    • textstring“ONE VOICE: the script to speak (≤900 characters)”
    • enginestring“voice-engine id: one voice defaults to 'seed-audio'; a dialogue defaults to the dialogue model ('eleven-dialogue'); or 'eleven-v4', 'minimax-speech', 'kokoro'……”
    • voicestring“ONE VOICE: a preset from the chosen engine (e.g.”
    • linesarray“DIALOGUE: the turns in speaking order”
    • dryRunboolean“true = the exact credits only, nothing made or charged”
  • generate_sfxon 1 other server

    “MAKE a sound effect or ambience that does not exist yet, from a description, as an MP3: a whoosh, a door slam, a cash-register ding, rain on a window, a busy cafe.”

    Publisher declares: read-only no · destructive no · idempotent no · open-world yes

    5 inputs, 1 required
    • describestringrequired“the sound in words: what makes it, where, how it moves, e.g.”
    • secondsnumber“length in seconds, 0.5-30 (default 5)”
    • loopboolean“true for a loopable bed”
    • enginestring“omit for 'seed-audio', or 'eleven-sfx'”
    • dryRunboolean“true = the exact credits only, nothing made or charged”
  • generate_musicon 15 other servers

    “RAW music: describe it (genre, mood, instruments, tempo), get an instrumental MP3.”

    Publisher declares: read-only no · destructive no · idempotent no · open-world yes

    3 inputs, 1 required
    • promptstringrequired“the music in words, e.g.”
    • secondsnumber“track length, up to 300”
    • modelstring“minimax-music-3 or lyria-3.5”
  • generate_texton 8 other servers

    “Text generation against the writing-model catalog (Claude, Gemini, GPT, Llama, DeepSeek…) — ad copy, hooks, scripts, rewrites, brainstorms.”

    Publisher declares: read-only no · destructive no · idempotent no · open-world yes

    3 inputs, 1 required
    • promptstringrequired“the writing task / question”
    • rawboolean“RAW MODEL ACCESS: send the prompt with NO Hermoso system prompt — the model answers as itself rather than as an ad copywriter.”
    • modelstring“a writing-model id from hermoso_capabilities (a Claude / Gemini / GPT / Llama / DeepSeek id) — omit for the default”
  • “Use this when the user asks for a video ad, Reel, TikTok ad, UGC video or commercial for their business or product It renders a finished ad on a real video model, on their brand (real product photos,…”

    Publisher declares: read-only no · destructive no · idempotent no · open-world yes

    23 inputs, 1 required
    • creativeobjectrequired“the FULL structured output of plan_ad (must contain video_storyboard)”
    • creatorstring“CAST A SAVED CREATOR in this ad — their id from list_creators, or the name you know them by (“Sarah”), or a PRESET AI creator from list_creators presets by exa…”
    • faceRoutestring“'face_lane' records the user's statement "I own the rights to this face" for the real person in their reference (paid plans; the likeness terms apply).”
    • modelstring“video model id from hermoso_capabilities (default: the engine the router picks for the plan: seedance-2.5 for every ad, a person talking to camera included).”
    • durationSecondsnumber“total ad length in seconds, 4–180 (clamped).”
    • aspectRatiostring“output aspect ratio, e.g.”
    • resolutionstring“Omit = the default tier, hermoso_capabilities defaultResolution (720p), the tier its `credits` prices are for.”
    • captionsboolean“burn the plan's per-scene on-screen words as caption pills.”
    • endCardboolean“append the branded end card.”
    • musicboolean or string“music bed: OFF unless asked.”
    • musicModelstring“compose the bed on minimax-music-3 or lyria-3.5, skipping the free library”
    • boardboolean“paint the storyboard board first (default: on for UGC and cinematic spots, off for a product-only commercial); false = render from the text shot list”
    • foundationsstring“product commercial / cinematic: 'choose' paints ONLY the foundations (identity sheet + four-up moodboard, or hero + location) and returns them, no video — then…”
    • moodboardPickinteger“which moodboard panel (1-4, left to right, top to bottom) the storyboard follows; default 1”
    • foundationImagesobject“images a foundations:'choose' call returned, reused as-is (not repainted)”
    • lockupboolean“brand wordmark + tagline composited over the closing seconds.”
    • textStylestring or object“THE LOOK of captions and the end card, only with captions or endCard and only when the user described one: the look in WORDS ("chunky yellow comic letters, pur…”
    • ttsVoicestring“voiceover voice name (e.g.”
    • dryRunboolean“return the routing decision (single pass vs stitched acts, resolved model + act lengths) and the estimated credits the real render reserves, WITHOUT submitting…”
    • allowGenericProductboolean“proceed even though this brand has NO product photo on file and the ad features a product — the packaging will be INVENTED.”
    • screensarray“real screenshots of the app or site for the shots that film it on a screen (URLs, upload_file URLs or Library ids), laid over them as cards”
    • screenChoicestring“the user's answer to needsAppScreens: brand = their saved app screens, site = found on their website (saved to the brand), without = none, device kept turned a…”
    • insertsarray“the user's own screens laid over the finished video at these seconds (each item's look says how), sound running underneath (about 2 credits).”
  • make_template_adon 1 other server

    “Use this when the user asks for a quick, low-cost ad or post built from their words and pictures: a slideshow, carousel, iMessage or ChatGPT chat, Notes, kinetic type, myth vs fact, an app tour, or t…”

    Publisher declares: read-only no · destructive no · idempotent no · open-world yes

    1 input, 1 required
    • configobjectrequired“MUST include config.template: 'custom' or a preset id, plus its fields; language ('Portuguese') for its own words.”
  • make_motion_filmon 1 other server

    “MOTION GRAPHICS and ANIMATED VIDEO in one call: kinetic typography (animated text), a logo animation or intro, a product launch film, an animated social cut.”

    Publisher declares: read-only no · destructive no · idempotent no · open-world yes

    23 inputs
    • recipestring“type, logo, launch, product or social; help:true lists the recipes”
    • hookstring“opening line; '*word*' = italic accent”
    • linesarray“more short lines, in order”
    • taglinestring“line shown under the headline or logo”
    • namestring“default: the saved brand's”
    • logostring“logo image: https, upload_file URL or Library id; default the brand's”
    • imagesarray“pictures, hero first (https, upload_file URL or Library id)”
    • sizestring“frame shape: 16:9, 9:16, 1:1 or 4:5”
    • secondsnumber“logo 4-10, others 6-60”
    • moodstring“music and pace”
    • musicstring“'off' = silent”
    • themestring“logo: dark for a light logo”
    • energynumber“how much the film keeps moving, 0 to 1; 0 films exactly what is written”
    • movingboolean“true: one thing physically happens in each picture, filmed by a video model.”
    • movesarray“per scene, what happens in its picture; omit and the planner writes them”
    • briefstring“what the film is about, for the planner”
    • modelstring“image-to-video model id from hermoso_capabilities; default is the best value for short shots”
    • resolutionstring“the moving shots: 720p or 1080p (default) where the model has them”
    • specobject“the whole film as JSON, from a quote or a saved film; help:true shows the spec language”
    • projectUrlstring“a past film's url; returns its saved spec to edit and render again”
    • helpboolean“true returns the spec language and recipes; renders nothing”
    • previewboolean“true renders a 1-credit contact sheet to approve before the full render”
    • confirmboolean“true renders the film for real credits; without it you get a free quote”
  • finish_videoon 1 other server

    “Post-process an EXISTING rendered video (its served mp4 URL) with the proven direct-response 'reviewer' finish and/or a film-grain pass — no AI model, ~30s, billed for its compute only.”

    Publisher declares: read-only no · destructive no · idempotent no · open-world yes

    7 inputs, 1 required
    • videoUrlstringrequired“the served URL of the video to finish (from a previous render/job); a URL, not a local file path; use upload_file for local files”
    • headerstring“header pill copy, ≤40 chars (required when pills is on)”
    • substring“accent sub-pill copy, ≤34 chars (usually the product/brand)”
    • pointsarray“3-4 proof points, ≤44 chars each”
    • accentstring“brand accent hex for the sub-pill”
    • pillsboolean“default true — set false for a grain-only pass”
    • grainboolean“default false — anti-AI film-grain finish”
  • “MECHANICAL post-production on an EXISTING video (URL): ordered primitives run by ffmpeg, ~2 credits flat, NO AI model, as a NEW video.”

    Publisher declares: read-only no · destructive no · idempotent no · open-world yes

    6 inputs, 2 required
    • videoUrlstringrequired“the video to edit: a render / Library URL, a direct file, or a public post link; a URL, not a local file path; use upload_file for local files”
    • opsarrayrequired“the ordered edit plan (max 6 ops)”
    • brandNamestring“override the workspace brand name”
    • domainstring“override the brand website”
    • accentstring“override the brand accent hex”
    • dryRunboolean“true = return the estimated credits this edit reserves and run nothing”
  • edit_timeline

    “Compose ANY edit or transition yourself; there is no preset list.”

    Publisher declares: read-only no · destructive no · idempotent no · open-world yes

    10 inputs, 1 required
    • videoUrlstring“the video 'this' refers to (optional when every segment names its own src)”
    • segmentsarrayrequired“the clips on the output, max 12, later ones on top.”
    • overlaysarray“HTML/CSS overlays animated with @keyframes, each shown between its start and end”
    • sfxarray“sound effects placed on the output timeline”
    • sizestring“canvas: '9:16' default, '1:1', '4:5', '16:9', any 'W:H' (1:3 to 3:1, on a 1080 short side) or exact 'WxH' pixels”
    • fpsnumber“the frame rate, a whole number from 24 to 60 (24, 25, 30, 60 are the usual ones).”
    • motionBlurboolean“default true: anything that moves gets a real shutter blur along its path”
    • reviewboolean“default true: a vision read of each seam (about 2 credits) comes back with the render, with the seam frames”
    • matchstring or object“every cut: 'auto' default”
    • reframestring or number“every cut: 'auto' default”
  • video_frameson 3 other servers

    “LOOK at a video's exact frames, free: one contact-sheet image of the frames from start to end at fps (e.g.”

    Publisher declares: read-only no · destructive no · idempotent no · open-world yes

    7 inputs, 1 required
    • urlstringrequired“the video”
    • startnumber“window start in seconds”
    • endnumber“window end in seconds”
    • fpsnumber“frames per second in the window (default ~16 frames across it)”
    • timesarray“exact seconds instead of a window (max 48; max 8 with save)”
    • saveboolean“true = save the frames at times / clips as full-size stills and return their URLs”
    • clipsarray“save: one still from the middle of each numbered clip (split at hard cuts; negative counts from the end)”
  • “Surgically re-render ONE time window (1.5-8s) of an existing rendered video and splice it back on the VIDEO TRACK ONLY — the rest of the video and ALL audio stay byte-identical.”

    Publisher declares: read-only no · destructive no · idempotent no · open-world yes

    11 inputs
    • videoUrlstring“the served URL of the master video to fix (not needed with `job`)”
    • startSecondsnumber“window start in seconds (not needed with `job`)”
    • endSecondsnumber“window end in seconds (window 1.5-8s; not needed with `job`)”
    • promptstring“what the replacement footage should show — describe the shot, matching the master's style.”
    • jobstring“a finished video-blocks film (explainer / faceless channel): its job id.”
    • blocknumber“with `job`: which block, 1-based (block 3 = 20-30s of a film of 10s blocks)”
    • atSecondsnumber“with `job`: the moment in the film instead of `block` ("the line at 0:12" = 12)”
    • linestring“with `job`: the new narration for that block, the exact words to say (said as written, never rewritten; refused for free when it cannot fit the block)”
    • refImagestring“optional product/style anchor image URL”
    • speechWindowsarray“[[start,end],...] windows with spoken lines — the fix window must not overlap these”
    • dryRunboolean“true = return the estimated credits this fix reserves and render nothing”
  • “Cut ONE long video into several RANKED, ready-to-post short clips (podcast, webinar, interview, talk, long ad cut -> Reels/Shorts/TikTok).”

    Publisher declares: read-only no · destructive no · idempotent no · open-world yes

    4 inputs, 1 required
    • videostringrequired“the long video to clip — a Vimeo/Loom/Dailymotion/Streamable/Rumble/Wistia/Twitch/TED watch URL, a direct https .mp4/.mov/.webm, or a Hermoso /generated/ URL; …”
    • countnumber“how many clips to cut, 1-8 (default 4)”
    • aspectRatiostring“clip shape: '9:16' (default), '1:1', '16:9', any 'W:H', or 'keep' for the source framing”
    • captionsboolean“burn subtitles into every clip.”
  • add_subtitleson 2 other servers

    “Burn subtitles into ANY video and get the .srt too; nothing is cut or re-rendered.”

    Publisher declares: read-only no · destructive no · idempotent no · open-world yes

    6 inputs, 1 required
    • videostringrequired“the video to subtitle; a URL, not a local file path; use upload_file for local files”
    • textStylestring or object“the look: words, a preset or fields; omit for the default”
    • burnboolean“false = only the .srt”
    • autostring“sentences (default) or words: timing unit for transcribed subtitles”
    • wordsPerCuenumber“words per subtitle line when auto is words”
    • cuesarray“your own lines in seconds, no overlap, max 90 chars, no emoji”
  • make_explaineron 1 other server

    “AI HOST EPISODE: format 'host_episode' = one presenter talking to camera, the camera changing every piece (frontal, three-quarter, close), the same host and set throughout, native voice, 16:9.”

    Publisher declares: read-only no · destructive no · idempotent no · open-world yes

    25 inputs
    • topicstring“what the explainer should teach or explain — a topic or a short brief (host_episode: or pass `script`)”
    • lanestring“'blocks' (default) = 10-second video blocks, real motion, one narrated line per block; 'stills' = the picture film (a still about every 1.5s, narrated, no vide…”
    • formatstring“blocks lane: 'explainer' (default; Gemini Omni 720p, 16:9 or 9:16, runs exactly the length asked) or 'faceless_channel' (a YouTube/TikTok faceless channel vide…”
    • creatorstring“host_episode: the host, a saved creator or a preset by name or id (list_creators)”
    • hostImagestring“host_episode: a photo URL of the host instead (upload_file for a local file); a real person's face needs a paid plan”
    • scriptstring“host_episode: the exact words, said verbatim and split at natural breaks into 4-30s pieces”
    • fromDraftstring“host_episode: a finished draft's job id, to render it in HD (720p) with the same script, cameras, set and host”
    • settingstring“host_episode: the set in words (default: written to fit the topic)”
    • camerasarray“host_episode: the rotation, ids frontal / three_quarter / close or framings in words; one entry = one fixed camera”
    • thumbnailboolean“host_episode: one thumbnail of the host (default true)”
    • durationSecondsnumber“target length 20-120s (default 60); drives the section count — ~10s of narration each, 3-8 sections”
    • frameDensitystring“how many pictures per second of narration, and therefore what it costs.”
    • aspectRatiostring“'9:16' default (a faceless channel video and a host episode default to 16:9)”
    • stylestring“visual style: 'cinematic' (default, photoreal); styled shortcuts editorial_collage, flat_vector, stickman, whiteboard, ink_marker, silhouette, storybook, paper…”
    • channelstring“the CHANNEL TYPE — it sets the pacing, the narration register and the default look, and is orthogonal to `style` (a named style always wins): explainer (casual…”
    • voicestring“narration voice name — omit for the default warm read”
    • captionsboolean“turn ON-SCREEN TEXT on.”
    • subtitlesboolean“which on-screen text, once `captions` is on.”
    • musicstring“music bed under the narration, measured to sit about 14 dB under the voice and sidechain-ducked beneath it.”
    • upscalenumber“optional FINAL upscale — 2 doubles each side, 4 quadruples.”
    • endCardboolean“append the branded end card.”
    • brandNamestring“brand name for the end card — omit to leave it unbranded”
    • dryRunboolean“true = return the estimated credits this explainer reserves (its own pricing, stopped at the hold) and render nothing.”
    • fromFilmstring“A PRESENTER ON A FINISHED FILM (only when the user asks for one): the job id of a finished video-blocks explainer or faceless-channel film.”
    • presenterobject“with fromFilm: who presents and how.”
  • “Finish a video-blocks explainer or faceless-channel film (make_explainer) that STOPPED partway: a block the video model would not render, or a server restart.”

    Publisher declares: read-only no · destructive no · idempotent no · open-world yes

    2 inputs, 1 required
    • jobIdstringrequired“the id of the film that stopped (job_...), as get_job or list_jobs shows it”
    • dryRunboolean“true = return the credits the resume reserves and render nothing”
  • “CONTENT FACTORY: turn ONE topic into a SERIES of faceless short videos (narration over pictures, no presenter), optionally in the STYLE of a trending reference video.”

    Publisher declares: read-only no · destructive no · idempotent no · open-world yes

    17 inputs
    • topicstring“what the series is about, in the user's words”
    • episodesnumber“how many episodes, 1-10 (default 1)”
    • episodeTitlesarray“your own episode titles, in order (sets the episode count)”
    • durationSecondsnumber“length of EACH episode, 20-120 seconds (default 30)”
    • referencestring“a video link whose style every episode copies”
    • aspectRatiostring“default: the reference's shape, else 9:16”
    • stylestring“override the look in words (e.g.”
    • captionsboolean“burn subtitles of every spoken word.”
    • voicestring“narrator: 'female' or 'male' (default follows the reference), or any narration voice name make_explainer takes”
    • formatstring“renderer lane: 'explainer' (default; exact length) or 'faceless_channel' (2K, whole 10s blocks)”
    • planOnlyboolean“write the episode plan and style card, render nothing (a small planning charge)”
    • plan“the `plan` a planOnly call returned, to run it without planning again”
    • moreOf“MORE EPISODES: the `plan` an earlier call returned.”
    • confirmboolean“REQUIRED true to start rendering (real spend).”
    • jobIdsarray“SCHEDULE MODE: the episode job ids to schedule (finished ones are booked)”
    • scheduleobject“SCHEDULE MODE: where and when the first episode posts; the rest follow every `everyHours` (default 24)”
    • postCaptionsarray“SCHEDULE MODE: one post caption per episode, in order (default: the episode title)”
  • “ONE-CLICK EFFECTS, the same ones as the /tools pages: a photo goes in, a styled picture comes out, and most can then become a 5-second clip.”

    Publisher declares: read-only no · destructive no · idempotent no · open-world yes

    9 inputs
    • effectstring“effect id from the list”
    • imagesarray“its pictures, in the order listed: URL, local path or Library id”
    • optionsobject“its listed options, e.g.”
    • textstring“a line printed exactly, on an effect that takes one”
    • describestring“your own words: added to the look, or the custom choice”
    • aspectRatiostring“one of its listed shapes”
    • stillstring“CLIP STEP: the finished picture to animate”
    • speechstring“CLIP STEP: a line said in the clip, on an effect that speaks”
    • confirmboolean“true = run it (real spend); without it, a free quote”
  • product_sizzleon 1 other server

    “Render an 18-30s music-led PRODUCT SIZZLE: ONE 15s Seedance 2.5 hero clip of the product, diced into fast cuts and intercut with typeset spec/CTA cards on a brand-coloured grain background, mixed to …”

    Publisher declares: read-only no · destructive no · idempotent no · open-world yes

    10 inputs, 1 required
    • promptstringrequired“what the sizzle should show — the product, the setting, the look”
    • secondsnumber“finished length, clamped to 18-30s (default 25).”
    • refImagestring“product packshot URL that anchors the real label — strongly recommended”
    • aspectRatiostring“'9:16' default; anything the Seedance row that renders it does not list falls back to 9:16”
    • resolutionstring“hero-clip resolution and therefore the whole cost — DEFAULT '1080p' (≈1,040 credits); '720p' ≈470, '480p' ≈220, '4k' ≈4,130”
    • specsarray“up to 4 spec lines for the typeset cards, ≤26 chars each”
    • ctastring“closing CTA line, ≤30 chars”
    • brandNamestring“brand name on the cards — defaults to the workspace brand”
    • musicMoodstring“music-bed mood, e.g.”
    • stylestring“the LOOK of the sizzle (same price for every look): hypermotion (Fast cuts, crash zooms and speed ramps), product_2d (The real product with playful hand-drawn …”
  • generate_videoon 32 other servers

    “Use this when the user asks for a video clip, b-roll or a shot from a description, or names a video model.”

    Publisher declares: read-only no · destructive no · idempotent no · open-world yes

    27 inputs
    • dryRunboolean“true = render NOTHING and return the credits this exact clip would hold (an estimate; the final charge is its actual cost).”
    • promptstring“the video prompt / shot description (for a refVideo edit, this is the transformation instruction); optional with `shots`”
    • rawboolean“Send the prompt to the model exactly as written, without the wording Hermoso usually adds (packaging and label guidance, reference lines, colour names for hex …”
    • refImagestring“local path or URL to anchor the first frame.”
    • speakstring“A PERSON SAYING THESE EXACT WORDS TO CAMERA — the default for any "make my photo talk" / spokesperson / talk-to-camera ask.”
    • creatorstring“STAR A SAVED CREATOR in this clip — their id from list_creators, or the name you know them by, or a PRESET AI creator from list_creators presets (exact name or…”
    • anglesboolean“OPTIONAL, OFF BY DEFAULT: with `creator`, also send the extra views that creator already has saved (their pose plates, up to 2) beside their portrait.”
    • refImagesarray“SEVERAL reference images (paths or URLs: a person, products, a place) that must all appear.”
    • refVideostring“URL of an existing video to EDIT rather than generate from scratch — the clip is transformed per your prompt, keeping the SOURCE clip’s aspect ratio and, on mo…”
    • endImagestring“local path or URL of the LAST frame: the clip travels from refImage (required with it) to this image.”
    • loopboolean“true = a seamless loop whose last frame flows back into its first.”
    • shotsarray“MULTI-SHOT: one clip cut into these shots, in order; the seconds add up to a length the model renders.”
    • extendboolean“true = EXTEND refVideo by durationSeconds (default 5) with what your prompt says happens next, as ONE clip.”
    • extendModestring“with extend: 'continue' = the SAME take past its last frame, only when asked seamless (models with extend in hermoso_capabilities).”
    • interactionIdstring“with extend:true on an Omni model: the interactionId returned by an earlier render on that model — continues it from its own stored context instead of re-uploa…”
    • durationSecondsnumber“length of THIS ONE clip in seconds: one of the CHOSEN model’s listed durations in hermoso_capabilities (they differ per model, 4–8s on the short ones up to 30s…”
    • aspectRatiostring“default '9:16'”
    • modelstring“video model id from hermoso_capabilities; a named model is never swapped without asking.”
    • faceRoutestring“'face_lane' records the user's statement "I own the rights to this face" for the real person in their reference (paid plans; the likeness terms apply).”
    • resolutionstring“Omit = the default tier, hermoso_capabilities defaultResolution (720p), the tier its `credits` prices are for.”
    • cameraMovestring“A named camera move around the still in refImage (orbit = a quarter turn, the default; orbit_half / orbit_full = a half turn / a turntable).”
    • cameraTrajectoryarray“Your own ordered camera path, 2 to 12 keyframes, for the camera-controls model only (same rule as cameraMove; overrides it).”
    • ttsScriptstring“voiceover script to speak”
    • ttsVoicestring“voice name, e.g.”
    • musicModelstring“compose the bed on minimax-music-3 or lyria-3.5, skipping the free library”

    and 2 more not listed here.

  • generate_avataron 5 other servers

    “ANIMATE A PHOTO so it talks: the presenter in `image` says `script` word for word, a separate voice lip-synced onto the still.”

    Publisher declares: read-only no · destructive no · idempotent no · open-world yes

    9 inputs, 2 required
    • imagestringrequired“local path or URL of the presenter portrait.”
    • scriptstringrequired“the words the avatar speaks”
    • voicestring“voice name (Aria / Sarah / George / Adam).”
    • enginestring“leave out for the standard engine; only an engine listed in hermoso_capabilities avatarEngines is accepted”
    • resolutionstring“'720p' (default) or '480p'”
    • promptstring“motion direction, only for an engine other than standard that hermoso_capabilities lists”
    • seedinteger“fixed seed, only for an engine other than standard”
    • acceptQueueboolean“only for an engine hermoso_capabilities marks oneAtATime: wait in line”
    • dryRunboolean“return the credits this exact job would hold, without rendering”
  • stitch_videoon 1 other server

    “Render a multi-scene STITCHED video (≥2 scenes) — ONLY for spots LONGER than ONE clip of the chosen model.”

    Publisher declares: read-only no · destructive no · idempotent no · open-world yes

    7 inputs, 1 required
    • scenesarrayrequired“array of scene objects (visual + optional voiceover/seconds)”
    • aspectRatiostring“output aspect ratio, e.g.”
    • voiceoverstring“full voiceover script spoken across the scenes”
    • voicestring“voiceover voice name, e.g.”
    • resolutionstring“720p (default), 1080p for full detail, or 480p for a cheaper draft”
    • modelstring“video model id from hermoso_capabilities — omit to let the router pick”
    • durationSecondsnumber“total spot length in seconds (defaults to the sum of the scenes’ seconds)”
  • “Poll a render job by id.”

    Publisher declares: read-only yes · destructive no · idempotent yes · open-world no

    1 input, 1 required
    • idstringrequired“the job id: starts with job_, from a render tool's jobId or from list_jobs (generate_voice and generate_sfx return their file directly and have no job)”
  • “Use this when the user wants a week of social posts planned, or answers yes or no to planning their week.”

    Publisher declares: read-only no · destructive no · idempotent no · open-world yes

    3 inputs
    • confirmboolean“true only after the user agreed to the price”
    • declineboolean“true when the user said no”
    • timezonestring“the user’s IANA timezone when known, e.g.”
  • “List the bundled Hermoso SKILLS — multi-step workflow instructions (SKILL.md) that orchestrate the other tools (research an ad space, plan+render a finished ad, product photoshoot, raw generation) — …”

    Publisher declares: read-only yes · destructive no · idempotent yes · open-world no

    Takes no inputs.

  • “Load a bundled skill’s full SKILL.md workflow instructions by name (from list_skills), or a built-in in-app skill by its id (e.g.”

    Publisher declares: read-only yes · destructive no · idempotent yes · open-world no

    1 input, 1 required
    • namestringrequired“bundle name from list_skills, e.g.”
  • “Save a reusable custom SKILL — a named creative directive/playbook applied to future ads (a hook formula, a UGC recipe, a compliance rule, a named specialist persona like “our founder-story style” or…”

    Publisher declares: read-only no · destructive no · idempotent no · open-world no

    3 inputs, 2 required
    • brandstring“profile id/name from list_brands, this call only”
    • namestringrequired“short skill name, e.g.”
    • directivestringrequired“the full instruction the skill applies when used (1–6 sentences, imperative)”
  • delete_skillon 5 other servers

    “Delete one of the workspace’s CUSTOM skills by id (from list_skills).”

    Publisher declares: read-only no · destructive yes · idempotent yes · open-world no

    2 inputs, 1 required
    • brandstring“profile id/name from list_brands, this call only”
    • idstringrequired“the custom skill id (from list_skills)”
  • “List the durable facts & preferences saved in this workspace’s Memory (what the studio remembers about the brand, audience, taste, and do/don’t rules) — the same Memory the web app shows.”

    Publisher declares: read-only yes · destructive no · idempotent yes · open-world no

    2 inputs
    • categorystring“filter to one bucket (Brand/Audience/Taste/Do/Don’t/Preference)”
    • limitnumber“max items (default 50, max 200)”
  • “Save a durable fact or PREFERENCE about the brand, audience, or the user’s creative TASTE (e.g.”

    Publisher declares: read-only no · destructive no · idempotent no · open-world no

    3 inputs, 1 required
    • brandstring“profile id/name from list_brands, this call only”
    • textstringrequired“the fact/preference, concise”
    • categorystring“short bucket: Brand, Audience, Taste, Do, Don’t, or Preference (default General)”
  • “Delete a saved Memory item by its id (from list_memory).”

    Publisher declares: read-only no · destructive yes · idempotent yes · open-world no

    2 inputs, 1 required
    • brandstring“profile id/name from list_brands, this call only”
    • idstringrequired“the memory item id (from list_memory)”
  • “Clean up and consolidate the workspace Memory: drops entries that are about how Hermoso, a tool, a connector or a platform API behaves (product behaviour, not the brand), drops phone numbers and emai…”

    Publisher declares: read-only no · destructive yes · idempotent yes · open-world no

    2 inputs
    • brandstring“profile id/name from list_brands, this call only”
    • confirmboolean“true to APPLY the proposal; omit to only see it”
  • list_swipefileon 3 other servers

    “List this workspace's SWIPEFILE — the saved-ad research board: every named collection and the ads/creatives kept in it (advertiser, headline, body copy, media URL, platform, when it was saved, and an…”

    Publisher declares: read-only yes · destructive no · idempotent yes · open-world no

    2 inputs
    • collectionstring“only list ads in this collection (by name or id) — omit for every collection”
    • limitnumber“max ads to return (default 50, max 500)”
  • save_to_swipefileon 3 other servers

    “Save one or more ads/creatives to a named SWIPEFILE collection, creating the collection if it does not exist — the headless twin of the heart on every ad card in the web app.”

    Publisher declares: read-only no · destructive no · idempotent yes · open-world no

    3 inputs, 2 required
    • brandstring“profile id/name from list_brands, this call only”
    • collectionstringrequired“the collection name — an existing one, or a new one to create”
    • itemsarrayrequired“the ads to save”
  • update_saved_creatoron 3 other servers

    “Set the OUTREACH STATUS and/or a NOTE on a creator already saved in the swipefile (find_creators -> save_to_swipefile, or the heart on a creator card).”

    Publisher declares: read-only no · destructive yes · idempotent yes · open-world no

    4 inputs, 1 required
    • brandstring“profile id/name from list_brands, this call only”
    • keystringrequired“the saved row's key from list_swipefile, e.g.”
    • statusstring“outreach status: new, contacted, replied, booked or passed”
    • notestring“replaces the existing note; pass "" to clear it”
  • list_playbookson 13 other servers

    “List the PLAYBOOKS saved in this workspace — the reusable strategy cards (winning hooks, angles, formats and the concrete plays to run) kept from teardowns, angle mining and creatives worth repeating.”

    Publisher declares: read-only yes · destructive no · idempotent yes · open-world no

    2 inputs
    • limitnumber“max playbooks to return (default 25, max 100)”
    • fullboolean“true to return every hook/angle/play in the text, not just the headline counts”
  • save_playbookon 3 other servers

    “Save a reusable PLAYBOOK — the strategy takeaways worth re-running: the hooks that work, the angles, the formats, and the concrete plays.”

    Publisher declares: read-only no · destructive no · idempotent no · open-world no

    7 inputs, 1 required
    • brandstring“profile id/name from list_brands, this call only”
    • namestringrequired“the playbook headline — what it is, in a few words”
    • hooksarray“the opening hooks worth reusing, verbatim”
    • anglesarray“the persuasion angles ({title, detail})”
    • formatsarray“the formats/recipes this plays best in (e.g.”
    • playsarray“the concrete plays to run ({title, detail}) — the actionable half”
    • sourcestring“where it came from, e.g.”
  • delete_playbookon 3 other servers

    “Delete a saved playbook by id (from list_playbooks).”

    Publisher declares: read-only no · destructive yes · idempotent yes · open-world no

    2 inputs, 1 required
    • brandstring“profile id/name from list_brands, this call only”
    • idstringrequired“the playbook id (from list_playbooks)”
  • list_creatorson 12 other servers

    “List this workspace’s SAVED CREATORS — the reusable on-camera cast (AI creators made here, a person pulled from a social profile, a consented photo upload).”

    Publisher declares: read-only yes · destructive no · idempotent yes · open-world no

    2 inputs
    • limitnumber“max creators to return (default 24)”
    • genderstring“filter the PRESET creators by gender (the saved cast is never filtered)”

Each quoted sentence is the first sentence of the publisher’s own description. The kind of tool and the “changes data” mark are our AI model’s reading of that description, not something we tested: two differently worded runs had to agree, otherwise no label is shown, and nobody has checked them by hand. A tool without the “changes data” mark is described as only reading, searching or calculating, or the two runs did not agree. No tool was called. Read the publisher’s full description before relying on it. “Publisher declares” lists the hints the server itself sends with the tool (MCP annotations: read-only, destructive, idempotent, open-world); they are the publisher's claims, not checked by us. Inputs are the parameters in the tool's own input schema: name, declared type, whether it is required, and the first sentence of its description. The first 100 of 188 tools are listed. Read 2026-11-08; remote servers are read the day they appear and again about every 30 days; package servers at the monthly refresh.

How well are its tools described?

Features #4 and #5

How clearly this server's own tool descriptions say what a tool does, and how consistent its tool list is. The first part is an AI model's yes/no reading of each description; the second part is counted. We do not turn it into a grade.

Descriptions

AI reading of each description

No AI reading was available for this server's tools (the two runs did not agree, or the listing was not reviewed).

Counted from the listing

Parameters with a description
100% of 957
Tools with a description
100% of 188
Typical description length
705 characters

The tool set

Counted from the listing

Tools
188
Size of the list
about 95,377 tokens for an agent to load (characters ÷ 4)
Naming
186 of 188 names use snake_case

AI reading: easy to mix up

No pair found.

The AI items cover 0 of 188 tools: for a tool where two differently worded runs disagreed, that item is left out. They were not checked by hand, and no tool was called. “Easy to mix up” means two tools whose names or descriptions overlap without saying how they differ; the AI found none, comparing tools within groups of 40. We do not rate whether the tool set is complete or whether side effects are spelled out, because our AI reading of those was not reliable enough.

How do I add it to my client?

Config from the registry entry

From the remote endpoint (Streamable HTTP)

We connected to this URL on 2026-11-08 (a read-only session: no tool was called).

Add to CursorAn install link in the format Cursor documents, holding the Cursor config below.

Claude Desktop or claude.ai
Settings → Connectors → Add custom connector, then paste:
https://app.hermoso.ai/mcp
Cursor · .cursor/mcp.json in a project, or ~/.cursor/mcp.json
{
  "mcpServers": {
    "hermoso-ai-ad-generator": {
      "url": "https://app.hermoso.ai/mcp"
    }
  }
}
VS Code · .vscode/mcp.json
{
  "servers": {
    "hermoso-ai-ad-generator": {
      "type": "http",
      "url": "https://app.hermoso.ai/mcp"
    }
  }
}
VS Code (command line) · Bash or zsh terminal (macOS, Linux)
code --add-mcp '{"name":"hermoso-ai-ad-generator","type":"http","url":"https://app.hermoso.ai/mcp"}'
Claude Code
claude mcp add --transport http hermoso-ai-ad-generator https://app.hermoso.ai/mcp
Cline · cline_mcp_settings.json (Cline CLI: ~/.cline/data/settings/; in the editor: MCP Servers → Configure → Configure MCP Servers)
{
  "mcpServers": {
    "hermoso-ai-ad-generator": {
      "type": "streamableHttp",
      "url": "https://app.hermoso.ai/mcp",
      "disabled": false,
      "autoApprove": []
    }
  }
}
Gemini CLI · ~/.gemini/settings.json, or .gemini/settings.json in a project
{
  "mcpServers": {
    "hermoso-ai-ad-generator": {
      "httpUrl": "https://app.hermoso.ai/mcp"
    }
  }
}
OpenAI Codex · ~/.codex/config.toml, or .codex/config.toml in a trusted project
[mcp_servers.hermoso-ai-ad-generator]
url = "https://app.hermoso.ai/mcp"
Zed · Zed settings file (command: zed: open settings file)
{
  "context_servers": {
    "hermoso-ai-ad-generator": {
      "url": "https://app.hermoso.ai/mcp"
    }
  }
}
Try it in the MCP Inspector · runs on your machine, opens in your browser · Node 22.19+
npx @modelcontextprotocol/inspector@2.11.0 --server-url https://app.hermoso.ai/mcp --transport http
  • The server runs on the operator's infrastructure; it may ask you to sign in when you first connect.
  • Zed's docs we read on 2026-10-26 describe remote servers by URL and do not name the transport.
  • The MCP Inspector lists the server's tools, resources and prompts; a tool you call there runs for real.

Built from the server's registry entry with each client's documented format (checked 2026-10-14; the Cursor link and the VS Code command 2026-10-19; the MCP Inspector command 2026-10-21; Cline, Gemini CLI, OpenAI Codex and Zed 2026-10-26). Values in < > are yours to fill in; this page never contains a key. Where the server needs a setting the entry does not declare, it can still fail to start.

Does the registry listing match what we found?

Feature #6

The registry checks an entry when it is published (each new version), and nothing re-checks it on a schedule. A difference below is about the listing, not a judgement of the publisher.

Registry saysWe foundResultSource · date
Remote endpoint app.hermoso.ai/mcpSession openedmatchesour test · 2026-11-08
No credential declared as requiredOpened a session without any credentialmatchesour test · 2026-11-08
Repository github.com/hermoso-ai/hermosoOpens publiclymatchesapi.github.com · 2026-11-07

Who made it?

Feature #7 · lineage

Facts about the publisher. We do not turn them into a trust score.

Namespace
io.github.hermoso-ai · 4 servers
GitHub owner
an organization (hermoso-ai)
Repository
github.com/hermoso-ai/hermoso
Stars
6
Repository age
91–180 days
Last push
0–30 days ago
Archived
no
Security policy
SECURITY.md

Source: the registry entry and api.github.com · 2026-11-07; security policy read from GitHub · 2026-11-07; read when the server is listed, then about every 30 days; an operator can ask for an earlier re-read.

Publish this server? Run these checks against it yourself before release, from your own machine: a server the command starts, or your own server at its URL after a one-time proof that you control it: npx protocolprobe-selftest.

Tool list changes

Our reads compared · no tool called

Hosted endpoint app.hermoso.ai/mcp

1 of 1 comparison between consecutive reads showed a change (2 reads, 2026-11-07 to 2026-11-08; version 1.0.0).

Last change seen between our reads on 2026-11-07 (version 1.0.0) and 2026-11-08 (version 1.0.0). Changed without a new version: both reads list version 1.0.0.

Changed (1):

edit_timeline: description

Compared from the tool lists our read-only connection recorded, separately for each hosted endpoint and each package: each tool's name, title, description, input schema and annotations (the order of keys and of tools is ignored). We see a list only when we read it: a change is shown between the two reads where we first saw it, and anything after 2026-11-08 is not known yet. No tool was called. How it is compared

Choosing

Record checks

Public records · not a safety rating

5 of 5 record checks met · 4 do not apply

  • Not ranked by downloads: a remote server publishes no download figure
  • Not ranked by stars: 6 stars (ranked from 10)
  • MetRegistry entry is active and says how to run itActive, with a package or a remote address.
  • MetSource repository link opensThe linked GitHub repository is public.
  • MetRepository is not archivedNot archived.
  • MetCode pushed to the repository in the 90 days before our checkLast push 0-30 days before our check.
  • MetRepository publishes a security policy (SECURITY.md)SECURITY.md found.
  • Does not applyPackage names its licence with a standard (SPDX) identifierRemote server: it ships no package.
  • Does not applyThe version the registry lists can be installed and is not withdrawnRemote server: it ships no package.
  • Does not applynpm package carries build provenance (attestation)Remote server: it ships no package.
  • Does not applyNo known critical or high advisory in the package or its dependenciesRemote server: it ships no package.

Read from public records only: the registry entry, npm/PyPI, GitHub and advisory databases, checked 2026-11-07 (security policy 2026-11-07). No tool was called. Only “met” and “not met” are counted, so the total differs between servers; a check that does not apply or that we could not run is listed, never counted against it. It is not a safety rating, and the same rules apply to every server. How each check works · Leaderboards · How we stay neutral

Common questions about io.github.hermoso-ai/hermoso-ai-ad-generator

Answered from the data above

Who publishes io.github.hermoso-ai/hermoso-ai-ad-generator?

The registry ties the namespace io.github.hermoso-ai to the GitHub account hermoso-ai; only its owner can publish under it. Its linked repository, github.com/hermoso-ai/hermoso, belongs to the same GitHub account. These are facts from the registry and GitHub, not a rating.

Is io.github.hermoso-ai/hermoso-ai-ad-generator safe to install?

We do not rate servers as safe or unsafe. 5 of its 5 applicable record checks are met (facts from public records, each listed on this page). What the records show: Its source repository opens publicly. Its repository had a push in the last 90 days. These are facts with their sources on this page, not a safety guarantee, and we did not run its tools.

Does io.github.hermoso-ai/hermoso-ai-ad-generator need an API key or sign-in?

It opened a session without any credential on 2026-11-08. The registry entry declares no required credential; no declaration is not a guarantee.

Has io.github.hermoso-ai/hermoso-ai-ad-generator been security tested?

Partly. 4 of the 15 checks that apply to it were exercised. It accepted a session without any credential and lists tools whose names suggest they change data. We did not call them. ProtocolProbe publishes no security verdict for a named server.

Does io.github.hermoso-ai/hermoso-ai-ad-generator publish a security policy?

Yes: its repository has a public security policy (github.com/hermoso-ai/hermoso/blob/HEAD/SECURITY.md). That is the publisher's statement, not permission for us to test the server, and we call none of its tools.

Where is the source code of io.github.hermoso-ai/hermoso-ai-ad-generator?

In github.com/hermoso-ai/hermoso, which opens publicly.

Found this page useful?

Anonymous · no sign-in

A quick, anonymous reaction — no account, no email. One per kind from your browser. The tally updates when the page is next refreshed, and it is only a rough count, never a safety rating or a measured number of people.

Compare with another server

Side by side: registry entry, our read-only session, the tools and the 18 classification fields. Same values as each page; no score.

Security testing

What did we test, and what did we leave out?

18 checks · features #1, #2

Open without sign-in

It accepted a session without any credential and lists tools whose names suggest they change data. We did not call them. No security score is given; a check we did not run is never counted as passed.

The 18 checks at a glanceour test · 2026-11-08
  1. Inventory: exercised
  2. Hidden instructions: exercised
  3. Real execution: not run · our policy
  4. Secret leak: not run · nothing to check it against
  5. Internal addresses: not run · our policy
  6. Unconfirmed changes: not run · our policy
  7. Bad input: not run · our policy
  8. Model requests: not run · nothing to check it against
  9. Sensitive questions: not run · nothing to check it against
  10. Folder boundary: not run · our policy
  11. Log leaks: not run · nothing to check it against
  12. Hidden content: exercised
  13. Fake assistant turns: not applicable
  14. Token passthrough: not applicable
  15. Token audience: not applicable
  16. Sign-in redirect: not run · our policy
  17. Consent handling: not testable from outside
  18. Sign-in metadata: exercised

4 exercised11 not run (reason given)3 not applicable

Each check ends in exactly one state: exercised, not run (with the reason), or not applicable (with the reason). 4 of the 15 checks that apply to it were exercised. We do not publish security-check outcomes for named servers. These 18 checks are separate from the record checks at the top of this page, which read public records only.

Exercised · 4
  • 1Inventory
    tool-list
    Reads the declared tools, resources and prompts.exercised
  • 2Hidden instructions
    description-poisoning
    Looks for hidden instructions in tool descriptions.exercised
  • 12Hidden content
    audience-hiding
    Checks resources that hide content from the user.exercised
  • 18Sign-in metadata
    oauth-url-scheme
    Checks the published sign-in metadata for unsafe URLs.exercised
Not run, with the reason · 11
  • 3Real execution
    real-execution
    Calls a tool and checks what really happens.This check calls the server's tools. ProtocolProbe calls no server's tools for anything it publishes, except for a server whose operator has claimed it and given permission (that server's page says so); an operator can run this check on their own server with our self-test command.not run · our policy
  • 4Secret leak
    canary-leak
    Plants a fake secret and checks whether it leaks.This check needs at least one successful tool call to have something to check, and no tool was called.not run · nothing to check it against
  • 5Internal addresses
    ssrf
    Checks whether a tool can be steered to internal addresses.This check calls the server's tools. ProtocolProbe calls no server's tools for anything it publishes, except for a server whose operator has claimed it and given permission (that server's page says so); an operator can run this check on their own server with our self-test command.not run · our policy
  • 6Unconfirmed changes
    write-without-confirmation
    Checks whether tools that change data act without asking.This check calls the server's tools. ProtocolProbe calls no server's tools for anything it publishes, except for a server whose operator has claimed it and given permission (that server's page says so); an operator can run this check on their own server with our self-test command.not run · our policy
  • 7Bad input
    error-handling
    Sends bad input and watches how the server answers.This check calls the server's tools. ProtocolProbe calls no server's tools for anything it publishes, except for a server whose operator has claimed it and given permission (that server's page says so); an operator can run this check on their own server with our self-test command.not run · our policy
  • 8Model requests
    sampling
    Checks whether the server asks the client's model to act for it.This check needs at least one successful tool call to have something to check, and no tool was called.not run · nothing to check it against
  • 9Sensitive questions
    elicitation-sensitive-data
    Checks whether the server asks users for sensitive data.This check needs at least one successful tool call to have something to check, and no tool was called.not run · nothing to check it against
  • 10Folder boundary
    root-boundary
    Checks whether the server reaches outside the folders it was given.This check calls the server's tools. ProtocolProbe calls no server's tools for anything it publishes, except for a server whose operator has claimed it and given permission (that server's page says so); an operator can run this check on their own server with our self-test command.not run · our policy
  • 11Log leaks
    logging-side-channel
    Checks whether log messages carry data they should not.This check needs at least one successful tool call to have something to check, and no tool was called.not run · nothing to check it against
  • 16Sign-in redirect
    open-redirect
    Checks the sign-in redirect for abuse.Testing the sign-in redirect means sending probes to the operator's sign-in service, which we do not do.not run · our policy
  • 17Consent handling
    confused-deputy
    Checks consent handling when one service acts for another.Depends on the operator's own consent screen, which differs per server and cannot be tested from outside.not testable from outside
Does not apply to this server · 3
  • 13Fake assistant turns
    fake-assistant-turn
    Checks prompts that pose as the assistant's own words.The server exposes no prompts.not applicable
  • 14Token passthrough
    token-passthrough
    Checks whether the server forwards a user's token to other services.The server opened a session without sign-in, so there is no token flow to check.not applicable
  • 15Token audience
    token-audience-validation
    Checks whether tokens meant for another service are accepted.The server opened a session without sign-in, so there is no token flow to check.not applicable

Source: our own test, 2026-11-08, mcp-runtime-engine@0.1.0. Which checks count and why follows our published scoring model.

What happened during the test?

Feature #8

The steps in order, as the test record holds them.

  1. 0 msStarted a read-only session with the server's remote endpoint
  2. 1,561 msConnected after 0 redirects. No credential was sent or needed.
  3. Read the inventory: 188 tools, 6 resources, 0 prompts.
  4. 9,658 msFinished: 4 checks exercised.

Reference

How is it classified?

18 fields

One value per field, each with its source. Fields written by AI models are marked. A field with no value says why. How often each is read again: methodology.

TopicAI-classified
pending: added at the next monthly refresh
listed 2026-11-07; not classified yet
Pricing modelAI-classified
pending: added at the next monthly refresh
the website is not read yet
Programming language
api.github.com · 2026-11-07 · read when a version is new, and at the monthly refresh
Hosting type
registry · 2026-11-07 · read from the registry entry; changes when a new version is listed
MCP capability
resources, tools
our test · 2026-11-08
Transport
streamable HTTP
registry · 2026-11-07 · read from the registry entry; changes when a new version is listed
Required credentials
none declared
registry · 2026-11-07 · no declaration is not a guarantee · read from the registry entry; changes when a new version is listed
Package registry
does not apply
remote-only server: no package is shipped, so there is no package registry
Freshness
0–30 days
registry updatedAt · 2026-11-07
Lifecycle status
registry · 2026-11-07 · status read from the registry every day
License class
does not apply
remote-only server: ships no package, so there is nothing to look up in a package registry
Integrity hash
does not apply
remote-only server: ships no package, so there is nothing to look up in a package registry
Signature / provenance
does not apply
remote-only server: ships no package, so there is nothing to look up in a package registry
Dependency advisories
does not apply
ships no package, so there are no dependencies to audit
Popularity
does not apply
remote-only server: ships no package, so there is nothing to look up in a package registry
Publisher signals
owner: an organization · repository age: 91–180 days · stars: 0–9
api.github.com · 2026-11-07 · new versions read the day they appear; the rest re-read about every 30 days
Maintenance
api.github.com · 2026-11-07 · new versions read the day they appear; the rest re-read about every 30 days
Environment
the publisher's infrastructure
where it runs, from the hosting type · 2026-11-07 · read from the registry entry; changes when a new version is listed

“Not stated” means the source did not state it; “does not apply” means it does not apply to this kind of server; “our limit” means our own lookup failed. A linked value opens the directory filtered to the same value.

Cite or correct this page

Sources and corrections

Suggested citation

ProtocolProbe. “io.github.hermoso-ai/hermoso-ai-ad-generator: MCP server record.” Data checked 2026-11-08. https://protocolprobe.com/mcp/servers/io.github.hermoso-ai/hermoso-ai-ad-generator

Figures are valid for the date shown. Our facts on this page are CC BY 4.0: reuse them with a link and the check date. Quoted text (the publisher’s descriptions, website quotes, tool names) is not ours to license.

As JSON: this record in the API · API documentation.

Badge for your README

ProtocolProbe: 5 of 5 record checks met

[![ProtocolProbe](https://protocolprobe.com/api/v1/badge/io.github.hermoso-ai/hermoso-ai-ad-generator)](https://protocolprobe.com/mcp/servers/io.github.hermoso-ai/hermoso-ai-ad-generator)

It shows this server's record-check count: facts from public records, no tool called. It is not a security rating or a pass. It links here and updates after each data reload.

Something wrong, or is this yours?

If a fact here is wrong, tell us and we will re-check it. If you operate this server, you can run our checks on it yourself with the free self-test command; we call no server’s tools ourselves for anything shown here, except for a server whose operator has claimed it and given permission (that server's page says so).

Write to hello@protocolprobe.com · corrections log · methodology