Escanor MCP
One MCP server for your infrastructure
Escanor MCP is a single Model Context Protocol server that sits in front of the integrations your workspace has connected, so a coding agent such as Claude Code needs one connection instead of dozens. Each call runs with your workspace's stored credential, and you can revoke an agent's key at any time.
What is MCP?
MCP (the Model Context Protocol) is how coding agents discover and call tools. Instead of configuring a separate server per provider, you point the agent at Escanor and it reaches every supported integration you connected in the dashboard.
01
Your coding agent
Claude Code, Cursor, VS Code, ChatGPT or any app that supports remote MCP servers.
02
Escanor MCP
Checks the agent’s key, finds the operation and runs it with your workspace’s stored credential.
03
Your providers
GitHub, Vercel, Sentry, Neon, Stripe and the other integrations you connected.
Five tools, thousands of operations
The cloud providers alone have thousands of operations, so an agent gets five tools instead of one per operation. It finds an operation with escanor_list_tools and runs it with escanor_invoke. If a provider is not connected, the answer says so plainly.
| Tool | Does |
|---|---|
| escanor_list_providers | Which providers exist. |
| escanor_connection_status | Which of them your workspace has connected, and which need reconnecting. |
| escanor_list_tools | The operations for one provider (searchable and paged). |
| escanor_usage_stats | Usage and rate limits. |
| escanor_invoke | Runs one operation by its id. |
Providers that run on your workspace connection
For these, connecting the integration in the dashboard is enough. More providers appear in the catalog with additional configuration requirements; see MCP providers for the full list.
- Source control
- GitHub, GitLab
- Deploy and hosting
- Vercel, Netlify, DigitalOcean, Terraform
- Data
- Neon, Supabase, Redis Cloud
- Monitoring and incidents
- Sentry, PagerDuty
- Work and communication
- Slack, Linear, Notion, Asana, DocuSign, SendGrid
- Payments
- Stripe
- AI
- OpenAI, Anthropic
Connect your coding agent
Open MCP in the dashboard. It creates a key and shows the exact steps for your app, already filled in with your address and key. The key is shown once.
- Claude Code
- Paste the one-line command the dashboard shows into your terminal.
- Claude Desktop
- Add the snippet to claude_desktop_config.json (it uses the mcp-remote helper) and restart.
- Cursor
- Paste the configuration into ~/.cursor/mcp.json or the MCP settings.
- VS Code
- Paste the configuration into .vscode/mcp.json, then use Copilot Chat in agent mode.
- ChatGPT
- Add a custom connector with the server address and sign in with your Escanor account. There is no key to copy.
- Anything else
- Any app that supports remote MCP servers can use the server address and an Authorization: Bearer header.
Credentials and safety
- Each call runs with your workspace's stored credential for that provider, only for the length of the call. Provider credentials stay on the server, never in prompts or tool arguments.
- Destructive operations need
confirm=truein their arguments, so nothing is deleted or overwritten by accident. - Read-only tokens and limited allowed-write policies keep each agent to the operations you choose. OIDC separates
mcp:readfrommcp:invoke. - Every key is listed with its last use; revoke one to cut off that agent immediately. MCP activity shows each call in plain language.
More in Permissions and approvals and API and keys.
See it in practice
Safe deploys for AI coding agents walks through a full setup, from the first read to an approved redeploy.