Openreview MCP server

@pipeworx/openreview

OpenReview MCP — venue submissions + peer reviews for ML conferences (ICLR, NeurIPS, ICML, COLM, EMNLP, etc.). API v2.

Part of Pipeworx — an MCP gateway connecting AI agents to 1679+ live data sources.

Tools

• list_venues(query?, limit?, offset?) — list active and historical venues
• getvenue(groupid) — group / venue metadata
• listsubmissions(venueid, query, limit?, offset?) — papers in a venue matching a topic
• get_note(id, details?) — individual note (paper, review, comment, decision)
• getpaper(forumid) — paper + all its threads (reviews, rebuttal, decision)
• searchnotes(query, contentfield?, signature?, venue_id?, limit?, offset?) — full-text search, optionally scoped to one venue

Upstream limitation: the /notes read path is challenged

Verified 2026-09-01. GET /notes?... on both api2.openreview.net and the v1 host answers 403 ChallengeRequiredError for every automated client — from Cloudflare egress and from an ordinary laptop alike, with any User-Agent, and with or without a token. /groups and /notes/search are unaffected.

What that means per tool:

No API key clears the challenge, so a failure here is deliberately not reported as a credential problem.

Auth

Public reads are keyless. Some endpoints (private invitations, restricted venues) need a token — optional:

• Platform key (optional): gateway env PLATFORMOPENREVIEWTOKEN
• BYO (optional): ?_apiKey= after logging in at openreview.net and grabbing a token

Data source

https://api2.openreview.net/ — v2 (also https://api.openreview.net/ v1 for legacy venues).

Quick Start

Add to your MCP client (Claude Desktop, Cursor, Windsurf, etc.):

{
  "mcpServers": {
    "openreview": {
      "url": "https://gateway.pipeworx.io/openreview/mcp"
    }
  }
}

What this endpoint actually serves

tools/list at https://gateway.pipeworx.io/openreview/mcp returns the tools in the table above plus the shared Pipeworx meta-tools — askpipeworx, discovertools, search_within, remember/recall and the rest of the gateway-wide set. So the tool count you see is larger than this table: a single-pack endpoint currently lists roughly 30 shared tools alongside the pack's own. The connection's initialize response states its exact scope, and is the authoritative answer for a given day.

This is deliberate, not multiplexing by accident. The meta-tools are what let a scoped connection answer a question this pack does not cover — via ask_pipeworx, which routes across the whole catalog — without you adding a second MCP server. There is currently no way to mount a pack endpoint without them; if the extra schemas cost you more context than the routing is worth, connect to the full gateway once rather than to several pack endpoints.

Or connect to the full Pipeworx gateway to get every pack's tools listed directly, instead of just this one's:

{
  "mcpServers": {
    "pipeworx": {
      "url": "https://gateway.pipeworx.io/mcp"
    }
  }
}

Both URLs reach the same gateway and the same 1679+ data sources. The only difference is which pack's tools are listed directly; ask_pipeworx reaches all of them from either one.

No MCP client? Call it over HTTP

curl -X POST https://gateway.pipeworx.io/v1/tools/list_venues \
  -H 'Content-Type: application/json' \
  -d '{"query":"ICLR"}'

Shortened. The full README is on GitHub.