The full upstream README, mirrored here for reference. Install config, tool schemas, adoption signals, and an original overview live on the AutoRFP.ai listing page.
The Model Context Protocol (MCP) server that lets AI assistants like Claude and ChatGPT connect directly to your AutoRFP.ai workspace. Ask questions about your RFP projects, search your approved content library, and check requirement status without leaving the chat.
This repo holds the public server.json manifest. The server itself is hosted by AutoRFP.ai.
Once connected, your AI assistant has read-only access to:
| What it can access | What it can do |
|---|---|
| Tags | See your organisation's tag categories and vocabulary |
| Projects | List projects, filter by status, due date, or tags |
| Requirements | Read questions and answers within a project, filter by status or tag |
| Content Library | Search approved content by keyword or semantic meaning, browse by tag or file type |
The connection cannot create, edit, or delete anything in AutoRFP.ai.
AutoRFP.ai runs three regional MCP endpoints. Pick the one that matches your AutoRFP.ai data residency region:
| Region | URL |
|---|---|
| APAC | https://api.autorfp.ai/mcp |
| EU | https://api.eu.autorfp.ai/mcp |
| US | https://api.us.autorfp.ai/mcp |
If you do not know your region, check the URL you log in with, or ask your AutoRFP.ai admin.
There are two parts: an org-level admin sets up the connection once, then each user authorises it for their own chat.
tags:read — see your organisation's tag categoriesprojects:read — read your RFP projects and requirementscontent:read — search your approved content libraryEach user inherits their own AutoRFP.ai permissions, so people only see what they would see logged into the app directly.
ChatGPT and any other MCP-compatible client uses the same server URLs and OAuth flow. Add a custom connector or integration in your client's settings and point it at the URL for your region.
Once connected, try:
"I don't see the Connectors option in Settings." You need a Claude Pro, Team, or Enterprise plan.
"A colleague added the integration but I still can't see my data." Each user connects their own AutoRFP.ai credentials. Run Part 2 of the setup above.
"I'm getting permission errors on certain projects." The connector inherits your AutoRFP.ai role-based permissions. If you cannot see a project in the app, the assistant cannot either.
"Which region am I on?" Look at the URL you use to log into AutoRFP.ai, or ask your AutoRFP.ai admin.
MIT