Gentkey MCP server
online · 96% uptimecom.gentkey/mcp·v0.1.0One MCP URL for all your connectors — scoped writes, enforced constraints, and a full audit trail.
0 stars
Reviews
Write oneNobody has reviewed Gentkey yet.
If you have run it, two minutes of your experience saves the next person an afternoon.
Gentkey tools
Tool list not cached yet. `describe` through the gateway fetches it live.
Public scan report
scanner v0.1.9 · 2026-09-24 · same rubric, same numbers if you re-run it
- –Code scanremote-only server, no package to scann/a
- Live reliabilityremote reachable in 252ms (auth required)20/20
- –Tool poisoningtools not inspected (endpoint requires auth); not countedn/a
- Auth qualityOAuth resource metadata advertised on 40115/15
- Maintenancelast push 28 days ago15/15
- Maintainer identitynamespace and repository owner differ; GitHub account older than a year; website matches verified namespace7/10
How Gentkey compares
Against every graded MCP server, from our own scans and uptime checks. Updated after each scan run.
- Safety score
- 95/100
- higher than 97% of 48,255 servers
- Uptime
- 96%
- of our recent checks; the typical remote server answers 100%
- Response time
- 519 ms
- faster than 28% of 30,031 servers
Install Gentkey in Claude Code, Cursor or VS Code
claude mcp add --transport http mcp https://app.gentkey.com/mcp
What the publisher says
From the Gentkey repository's README, as published. We do not edit it. Read it on GitHub
Gentkey
One MCP URL for all your connectors.
Connect your accounts once at gentkey.com. Every AI surface — claude.ai (web & mobile), Claude Code, Cursor, any MCP client — reaches all of them through a single OAuth-protected MCP endpoint:
https://app.gentkey.com/mcpProviders give an agent a key to the whole house. Gentkey cuts it a key to one room — and you can take it back at any time.
Why
Connectors are a commodity; turning a coarse OAuth token into scoped, revocable, auditable capabilities is the point:
- Capability-scoped writes. Read tools work out of the box (switchable off per connection). Write tools require an explicit capability grant you make in the dashboard.
- Enforced constraints. Grants are bounded per call (e.g. max $50 budget delta) and per rolling window (writes / targets / dollars per hour or day).
- Dry runs. Every write tool advertises a dry_run flag that reports exactly what the real call would do — including the policy verdict, even before a grant exists — with no side effects.
- Full audit trail. Every decision is recorded (allowed / denied / constraint-denied / dry-run) with a human-readable plan summary — "Change 'Summer Sale' budget: $50 → $80", not grant:abc.
- Server-side credentials. Upstream credentials are encrypted at rest (AES-256-GCM, AAD-bound; optional Cloud KMS envelope encryption) and injected server-side. The model never sees them.
Connect a client
Create an account at app.gentkey.com, connect your services, then point your client at the endpoint. Authentication is OAuth 2.0 with dynamic client registration — no API keys to paste.
claude.ai (web & mobile) — Settings → Connectors → Add custom connector → https://app.gentkey.com/mcp
Claude Code
claude mcp add --transport http gentkey https://app.gentkey.com/mcpCursor — add to ~/.cursor/mcp.json:
{ "mcpServers": { "gentkey": { "url": "https://app.gentkey.com/mcp" } } }Any other client that speaks Streamable HTTP with OAuth works the same way.
What you can connect
- Native connectors with deep, semantic constraints: Google Ads (operation-typed capability grants, dollar caps on budget changes), Meta Ads (currency-aware budget deltas, dry-run plan/commit), BigQuery (statement-typed grants), Apple Search Ads, App Store Connect, Google Search Console, RevenueCat, and Superwall (per-resource capability maps).
- A directory of 1,798 services reachable by MCP passthrough, OpenAPI, or GraphQL — Notion, Linear, Slack, Stripe, Figma, Canva, Atlassian, HubSpot, Supabase, Sentry, and more. Upstream tools are proxied verbatim (schemas included) but default to write-gated: only curated read tools run without a grant, and every call is audited either way.
Full list with per-connector grant taxonomies: docs/connectors.md.
About this repository
Gentkey's source is proprietary. This repository hosts the public documentation, the MCP registry manifest (server.json), and the issue tracker — bug reports, questions, and connector requests are welcome in Issues.
- Connector catalog & grant taxonomies
- Security model
- FAQ
- Website: gentkey.com
- App: app.gentkey.com
- MCP endpoint: https://app.gentkey.com/mcp (Streamable HTTP, OAuth 2.0 + dynamic client registration)
Nothing above is checked by us. What we check is on the safety report.
Gentkey: common questions
- Is Gentkey MCP server safe?
- Yes, by our scan: it is graded A (95/100). Read the Gentkey safety report
- How do I install Gentkey?
- It runs remotely at app.gentkey.com. Add it to Claude Code, Cursor, VS Code or Claude Desktop with the snippets above, or call it through the mcp.market gateway without installing anything.
- Does Gentkey need an API key?
- No key to paste: it signs you in with OAuth when your client connects.
- Is Gentkey maintained?
- The last commit was 69 days ago (2026-08-27). The latest release is v0.1.0.
- Is Gentkey up?
- 96% of our last 28 checks got an answer. We check remote servers about four times a day.
- What can I use instead of Gentkey?
- Servers from other publishers that do the same job: SSH Manager MCP server and Tradememory Protocol MCP server.
- How does Gentkey compare with other MCP servers?
- safety score 95/100 (higher than 97% of 48,255 servers); uptime 96% (of our recent checks; the typical remote server answers 100%); response time 519 ms (faster than 28% of 30,031 servers). Full comparison.
Alternatives to Gentkey
Same job from other publishers: the closest match first, then the best rated.
- SSH ManagerSSH for AI agents: commands, files, backups and health, with approval, a vault and an audit trailnot reviewedEstablishedA
- Tradememory ProtocolTamper-evident decision audit trail and outcome-weighted memory for AI trading agents.not reviewedEstablishedB