Citation Velocity
Agent access

Built for AI agents first

Citation Velocity tracks which research papers are gaining citations fastest, across bioRxiv, arXiv, PubMed, Hugging Face Papers, and OpenReview. Agents connect through a native MCP server or the REST API and get structured, citable results: every paper carries a DOI link and a ready-to-paste citation. Use it for literature scans, innovation signals, and briefs that need current sources.

1. Get an API key
Keys belong to a human account and can be revoked at any time

1. Sign in with your email (Magic.link, no password). 2. Open Settings and create an agent API key: read scope for research, add write scope only if the agent should refresh the corpus. 3. Give the key to your agent as a Bearer token. The key is shown once; only its hash is stored here.

2. Connect the MCP server
Streamable HTTP at https://your-deployment.example/api/mcp

Claude Code

Muse mcp add --transport http citation-velocity https://your-deployment.example/api/mcp \
  --header "Authorization: Bearer cv_live_YOUR_KEY"

Claude Desktop, Cursor, and other MCP clients (JSON config)

{
  "mcpServers": {
    "citation-velocity": {
      "url": "https://your-deployment.example/api/mcp",
      "headers": {
        "Authorization": "Bearer cv_live_YOUR_KEY"
      }
    }
  }
}

ChatGPT: enable Developer Mode, add a connector with the MCP URL above and the same Bearer key as an Authorization header. Discovery card: /.well-known/mcp.json

Raw JSON-RPC check

curl -X POST https://your-deployment.example/api/mcp \
  -H "Authorization: Bearer cv_live_YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{"jsonrpc":"2.0","id":1,"method":"tools/call","params":{"name":"get_trending_papers","arguments":{"limit":5}}}'
Tools
One data layer powers both MCP and REST, so answers match everywhere
search_papersreadKeyword and filter search across the corpus, live OpenAlex, Hugging Face Papers, and OpenReview, ranked by citation velocity.
get_paperreadFull detail on one DOI: abstract, velocity, cross-domain spread, Altmetric attention.
get_trending_papersreadThe highest citation velocity papers right now.
get_paper_citationsreadWho is citing a paper, and which domains they come from.
get_alertsreadVelocity spike and cross-domain signal alerts, newest first.
get_statsreadCorpus size, top domain flows, top movers, last ingestion.
get_talent_radarreadResearchers ranked by Commercial Gravity Score.
get_founders_tracereadCached founder and commercialization signals for a paper's authors.
ingest_paperswriteRefresh the corpus from OpenAlex, Hugging Face Papers, and OpenReview, and regenerate signals and alerts.

Prompts

research_brief: Citation-backed brief on a topic, seeded with live search results.

innovation_scan: Breakthrough scan of one domain, seeded with its trending papers.

paper_deep_dive: Full assessment of one paper, seeded with its complete record.

Resources: citation://papers/trending, citation://stats, citation://alerts/recent, citation://papers/{doi}

Or use the REST API
OpenAPI 3.1 spec at /api/openapi.json and the agent guide at /llms.txt

Search papers

curl "https://your-deployment.example/api/v1/search?q=CRISPR+gene+therapy&limit=5" \
  -H "Authorization: Bearer cv_live_YOUR_KEY"

GET /api/v1/papers list and filter (q, domain, source, signal, min_velocity, sort, limit)

GET /api/v1/papers/{doi} one paper; add ?format=md for markdown

GET /api/v1/citations/{doi} citing papers and cross-domain spread

GET /api/v1/trending, /api/v1/stats, /api/v1/alerts, /api/v1/talent

POST /api/v1/ingest refresh the corpus (write scope)

Citation rules for agents

Cite papers by DOI link and use each record's cite_as string in bibliographies. Velocity and signal claims come from Citation Velocity; bibliographic data comes from OpenAlex. Citation velocity is citations per week since publication. Citations from 3 or more outside domains count as a cross-domain breakthrough signal.