MCP servers › com.synapseorch › synapseorch
MCP server · registry snapshot 2026-11-17 · last checked 2026-11-18
synapseorch MCP server
Registry name com.synapseorch/synapseorch
by synapseorch.com · domain (the registry's namespace rule)repository owner on GitHub: synapseorch-ai (organization)Source ↗Website ↗
Build and run AI agent workflows: start runs, answer what they ask, search knowledge, edit agents. — the publisher's description
Summary
com.synapseorch/synapseorch is a remote MCP server listed in the official MCP registry at version 2.0.12, under the namespace of the domain synapseorch.com. Its linked repository is github.com/synapseorch-ai/synapse-ai; GitHub reports its owner as synapseorch-ai, an organization.
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- not applicable: The version the registry lists can be installed and is not withdrawn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
- #466 of 28,859 repositories by GitHub stars, tied with 1 other
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
- required
- Repository
- public
- Last tested
- 2026-11-17
How a client gets in
OAuth sign-in; a client registration endpoint is published
Read from the server's public metadata on 2026-11-17 (api.synapseorch.com), in our read-only connection; no tool was called. It may have changed since. How it is read
Tool list
Not read: The server required a sign-in before it would list anything, and we never use a credential its operator did not give us. Its operator can test it with their own credential, on their own machine, with our self-test command (/docs/selftest). So there is no tool-list size and no result from the three APIs.
How fresh this is
- Live-tested
- attempted, see below
- Last tested
- 2026-11-17
- Latest version
- 2.0.12, published 2026-11-16 (registry, as of the 2026-11-17 snapshot)
- First listed version
- 2026-11-12 (earliest version still in the registry; older removed versions would be earlier)
- Version history
- 4 versions listed, 2026-11-12 to 2026-11-16; 4 in the 90 days before the 2026-11-17 snapshot.(versions still listed in the registry only)
Most recent 4 versions
2.0.12· 2026-11-162.0.11· 2026-11-152.0.8· 2026-11-132.0.6· 2026-11-12
- Registry data
- snapshot 2026-11-17
- Package and repository
- checked 2026-11-18
- 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
Version 2.0.11 (published 2026-11-15) → 2.0.12 (published 2026-11-16), between the registry lists of 2026-11-16 and 2026-11-17.
The publisher’s GitHub release note for v2.0.12, published 2026-11-15.
Read the release note (the publisher’s text, shown as plain text)
# Synapse 2.0.12 Synapse AI asks before it guesses and says what it assumed, a new request typed beside an open workflow is built as new, and Home never says nothing is waiting before it knows. ## Changed - **Synapse AI asks what it can't know, at most twice, and says what it assumed.** When a request leaves something only you can say, such as which terms a document is checked against, Synapse AI asks, in no more than two rounds of questions. Past that it builds, and the plan card lists what it assumed under **What I assumed**. Every card now says **How it starts** (a form, a schedule, or by hand), and a part of your request that only one branch of the workflow would do is called out. - **A new request typed beside an open workflow is built as new.** When a workflow is open and you describe a different job, Synapse AI builds it as a new workflow instead of reading it as a change to the open one. When it can't tell, it asks once whether you mean a change or something new. "Run it" right after a request that built nothing now says so and offers to build it. - **No stronger model is suggested before a build unless it is measured to be better.** Synapse AI used to suggest a bigger model before many builds, even when you had picked your model yourself. It no longer does. When it offers to try again on a stronger model after a failure, it names both models by their names. - **Saving to your workspace's own Vault needs no approval step.** A plan that writes to the Vault no longer needs a person to approve that step first. Sending email, posting to Slack, calling outside services and your own custom tools still ask first, as before. - **A model CLI that is signed out is no longer offered.** When Claude Code, Gemini CLI, Codex or GitHub Copilot's CLI is installed but not signed in, the model pickers leave its models out. A model you already saved on it stays visible where it is set. ## Fixed - **The OpenAI Agent Builder import asks for the export.** Sending the import line with nothing pasted under it no longer builds a guess. The panel holds the line back and says where the export comes from (in Agent Builder: Code → Agents SDK), in a note you can read in full. - **Synapse AI explains what it can build instead of a dead end.** When a very long request left Synapse AI's planner answering in prose, the turn could end with "I couldn't turn that into something I can make or change here." It now explains what can be built here and offers choices. - **Home never says nothing is waiting before it knows.** While Home was still loading, or when something couldn't be read, it could say "Nothing is waiting on you.", "No runs yet" or "Nothing runs on its own yet." Each part of Home now waits for what it reads, and if something can't be read, Home says so, with **Try again**, instead of showing an empty list or a zero. ## Install **macOS and Linux** ```bash curl -LsSf https://synapseorch.com/install.sh | bash ``` **Windows (PowerShell)** ```powershell irm https://synapseorch.com/install.ps1 | iex ``` <details> <summary>npm, pip or Docker</summary> ```bash npm install -g synapse-orch-ai && synapse pip install synapse-orch-ai && synapse # Python 3.11–3.14 and Node.js 20.9+ ``` For Docker, see [Docker](https://docs.synapseorch.com/getting-started/docker): the image is `synapseorchai/synapse-ai:2.0.12`, also `:latest`. </details> ## Upgrading - **From 2.0.x:** run `synapse upgrade`. Your data stays where it is, and the database is backed up first. - **From 1.x:** run `synapse upgrade` on your 1.x install. It moves you to 2.0 the same way 2.0.0 did; read the [2.0.0 notes](https://github.com/synapseorch-ai/synapse-ai/releases/tag/v2.0.0) for what changes. **Full changelog:** [v2.0.11...v2.0.12](https://github.com/synapseorch-ai/synapse-ai/compare/v2.0.11...v2.0.12) · [Documentation](https://docs.synapseorch.com)
Source: the official MCP Registry list on 2026-11-16 and on 2026-11-17; release notes from the GitHub Releases API, checked 2026-11-17.
What it is
How do I add it to my client?
Config from the registry entry
From the remote endpoint (Streamable HTTP)
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://api.synapseorch.com/api/v1/mcp
Cursor · .cursor/mcp.json in a project, or ~/.cursor/mcp.json
{
"mcpServers": {
"synapseorch": {
"url": "https://api.synapseorch.com/api/v1/mcp"
}
}
}VS Code · .vscode/mcp.json
{
"servers": {
"synapseorch": {
"type": "http",
"url": "https://api.synapseorch.com/api/v1/mcp"
}
}
}VS Code (command line) · Bash or zsh terminal (macOS, Linux)
code --add-mcp '{"name":"synapseorch","type":"http","url":"https://api.synapseorch.com/api/v1/mcp"}'Claude Code
claude mcp add --transport http synapseorch https://api.synapseorch.com/api/v1/mcp
Cline · cline_mcp_settings.json (Cline CLI: ~/.cline/data/settings/; in the editor: MCP Servers → Configure → Configure MCP Servers)
{
"mcpServers": {
"synapseorch": {
"type": "streamableHttp",
"url": "https://api.synapseorch.com/api/v1/mcp",
"disabled": false,
"autoApprove": []
}
}
}Gemini CLI · ~/.gemini/settings.json, or .gemini/settings.json in a project
{
"mcpServers": {
"synapseorch": {
"httpUrl": "https://api.synapseorch.com/api/v1/mcp"
}
}
}OpenAI Codex · ~/.codex/config.toml, or .codex/config.toml in a trusted project
[mcp_servers.synapseorch] url = "https://api.synapseorch.com/api/v1/mcp"
Zed · Zed settings file (command: zed: open settings file)
{
"context_servers": {
"synapseorch": {
"url": "https://api.synapseorch.com/api/v1/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://api.synapseorch.com/api/v1/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.
What do the published files and outside lists say?
Read as text · nothing run · no tool called
Package files
- MCP SDK in the package files. Not read yet: added with the next daily run.
- Build workflow. Not read yet: added with the next daily run.
Outside lists
- CVE records (NVD). No CVE record in the NVD feeds we read links this server's repository. Source: read 2026-11-17. This product uses the NVD API but is not endorsed or certified by the NVD.
- GitHub advisories. No GitHub advisory names this entry's listed package. Source: read 2026-11-17. GitHub Advisory Database, CC BY 4.0
- Docker MCP catalogue. Not in the Docker MCP catalogue on 2026-11-17. Source: read 2026-11-17.
- Cline marketplace. Not in the Cline marketplace on 2026-11-17. Source: read 2026-11-17.
Own host
- security.txt. No security.txt at https://api.synapseorch.com/.well-known/ (HTTP 404, 2026-11-17). No security.txt at https://synapseorch.com/.well-known/ (HTTP 404, 2026-11-17). Source: api.synapseorch.com/.well-known/security.txt · read 2026-11-17.
- A2A agent card. No A2A agent card at https://api.synapseorch.com/.well-known/ (HTTP 404, 2026-11-17). No A2A agent card at https://synapseorch.com/.well-known/ (HTTP 404, 2026-11-17). Source: api.synapseorch.com/.well-known/agent-card.json · read 2026-11-17.
Read as text from published files and outside lists. Nothing was installed or run, and no tool was called. A fact, not a rating. How each fact is read
Does the registry listing match what we found?
Feature #6The 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 says | We found | Result | Source · date |
|---|---|---|---|
| Remote endpoint api.synapseorch.com/api/v1/mcp | Endpoint answered and asked for sign-in | matches | our test · 2026-11-17 |
| No credential declared as required | Required sign-in | differs | our test · 2026-11-17 |
| Repository github.com/synapseorch-ai/synapse-ai | Opens publicly | matches | api.github.com · 2026-11-18 |
Who made it?
Feature #7 · lineage
Facts about the publisher. We do not turn them into a trust score.
- Namespace
- com.synapseorch
- GitHub owner
- an organization (synapseorch-ai)
- Repository
- github.com/synapseorch-ai/synapse-ai
- Stars
- 343
- Repository age
- 181–365 days
- Last push
- 0–30 days ago
- Archived
- no
- Security policy
- SECURITY.md
Source: the registry entry and api.github.com · 2026-11-18; security policy read from GitHub · 2026-11-12.
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.
Choosing
Record checks
Public records · not a safety rating5 of 5 record checks met · 4 do not apply
- Not ranked by downloads: a remote server publishes no download figure
- #466 of 28,859 repositories by GitHub stars, tied with 1 other
- 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 2026-11-18 (security policy 2026-11-12). 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 com.synapseorch/synapseorch
Answered from the data aboveWho publishes com.synapseorch/synapseorch?
The registry ties the namespace com.synapseorch to the domain synapseorch.com; only its owner can publish under it. Its linked repository is github.com/synapseorch-ai/synapse-ai; GitHub reports its owner as synapseorch-ai, an organization. These are facts from the registry and GitHub, not a rating.
Is com.synapseorch/synapseorch 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 com.synapseorch/synapseorch need an API key or sign-in?
It required sign-in on 2026-11-17. The registry entry declares no required credential; no declaration is not a guarantee.
Has com.synapseorch/synapseorch been security tested?
Partly. 1 of the 18 checks that apply to it was exercised. The server requires sign-in. We never use a credential its operator did not give us, so most checks could not run. ProtocolProbe publishes no security verdict for a named server.
Does com.synapseorch/synapseorch publish a security policy?
Yes: its repository has a public security policy (github.com/synapseorch-ai/synapse-ai/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 com.synapseorch/synapseorch?
In github.com/synapseorch-ai/synapse-ai, which opens publicly.
Found this page useful?
Anonymous · no sign-inA 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
Needs the operator's credential
The server requires sign-in. We never use a credential its operator did not give us, so most checks could not run. No security score is given; a check we did not run is never counted as passed.
Why no live read-only session: The server required a sign-in before it would list anything, and we never use a credential its operator did not give us. Its operator can test it with their own credential, on their own machine, with our self-test command (/docs/selftest). — the server's own choice
- Inventory: not run · needs the operator's credential
- Hidden instructions: not run · needs the operator's credential
- Real execution: not run · needs the operator's credential
- Secret leak: not run · needs the operator's credential
- Internal addresses: not run · needs the operator's credential
- Unconfirmed changes: not run · needs the operator's credential
- Bad input: not run · needs the operator's credential
- Model requests: not run · needs the operator's credential
- Sensitive questions: not run · needs the operator's credential
- Folder boundary: not run · needs the operator's credential
- Log leaks: not run · needs the operator's credential
- Hidden content: not run · needs the operator's credential
- Fake assistant turns: not run · needs the operator's credential
- Token passthrough: not run · needs the operator's credential
- Token audience: not run · needs the operator's credential
- Sign-in redirect: not run · our policy
- Consent handling: not testable from outside
- Sign-in metadata: exercised
1 exercised17 not run (reason given)0 not applicable
Each check ends in exactly one state: exercised, not run (with the reason), or not applicable (with the reason). 1 of the 18 checks that apply to it was 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.
- 18Sign-in metadata
oauth-url-schemeChecks the published sign-in metadata for unsafe URLs.exercised
- 1Inventory
tool-listReads the declared tools, resources and prompts.The server requires sign-in, and we never use a credential its operator did not give us.not run · needs the operator's credential - 2Hidden instructions
description-poisoningLooks for hidden instructions in tool descriptions.The server requires sign-in, and we never use a credential its operator did not give us.not run · needs the operator's credential - 3Real execution
real-executionCalls a tool and checks what really happens.The server requires sign-in, and we never use a credential its operator did not give us.not run · needs the operator's credential - 4Secret leak
canary-leakPlants a fake secret and checks whether it leaks.The server requires sign-in, and we never use a credential its operator did not give us.not run · needs the operator's credential - 5Internal addresses
ssrfChecks whether a tool can be steered to internal addresses.The server requires sign-in, and we never use a credential its operator did not give us.not run · needs the operator's credential - 6Unconfirmed changes
write-without-confirmationChecks whether tools that change data act without asking.The server requires sign-in, and we never use a credential its operator did not give us.not run · needs the operator's credential - 7Bad input
error-handlingSends bad input and watches how the server answers.The server requires sign-in, and we never use a credential its operator did not give us.not run · needs the operator's credential - 8Model requests
samplingChecks whether the server asks the client's model to act for it.The server requires sign-in, and we never use a credential its operator did not give us.not run · needs the operator's credential - 9Sensitive questions
elicitation-sensitive-dataChecks whether the server asks users for sensitive data.The server requires sign-in, and we never use a credential its operator did not give us.not run · needs the operator's credential - 10Folder boundary
root-boundaryChecks whether the server reaches outside the folders it was given.The server requires sign-in, and we never use a credential its operator did not give us.not run · needs the operator's credential - 11Log leaks
logging-side-channelChecks whether log messages carry data they should not.The server requires sign-in, and we never use a credential its operator did not give us.not run · needs the operator's credential - 12Hidden content
audience-hidingChecks resources that hide content from the user.The server requires sign-in, and we never use a credential its operator did not give us.not run · needs the operator's credential - 13Fake assistant turns
fake-assistant-turnChecks prompts that pose as the assistant's own words.The server requires sign-in, and we never use a credential its operator did not give us.not run · needs the operator's credential - 14Token passthrough
token-passthroughChecks whether the server forwards a user's token to other services.Needs a real token issued by the server's own sign-in service, which only its operator can give us.not run · needs the operator's credential - 15Token audience
token-audience-validationChecks whether tokens meant for another service are accepted.Needs a real token issued by the server's own sign-in service, which only its operator can give us.not run · needs the operator's credential - 16Sign-in redirect
open-redirectChecks 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-deputyChecks 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
Source: our own test, 2026-11-17, 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.
- 0 msStarted a session with the server's remote endpoint
- The server asked for sign-in before listing anything.
- 2,888 msFinished: 1 check 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.
“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. “com.synapseorch/synapseorch: MCP server record.” Data checked 2026-11-18. https://protocolprobe.com/mcp/servers/com.synapseorch/synapseorch
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
[](https://protocolprobe.com/mcp/servers/com.synapseorch/synapseorch)
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