Asta Scientific Corpus Search

Access Allen AI's massive scientific literature graph: 225M+ papers, 80M+ authors, 2.4B+ citation edges, and 12M+ full-text publications (285M+ passages).

When to Use

• Always as part of multi-source literature search (Channel 4 alongside PubMed, OpenAlex, Semantic Scholar)
• Especially useful when you need:
  • Full-text paragraph-level search (not just title/abstract)
  • Deep citation network traversal
  • Cross-disciplinary paper discovery
  • Author relationship analysis

API Configuration

Parameter	Value
Endpoint	https://asta-tools.allen.ai/mcp/v1
Protocol	MCP over HTTP POST (JSON-RPC style)
Auth	x-api-key: $ASTA_API_KEY header
Rate limits	Higher with API key; basic access without
Key registration	Free at https://allenai.org/asta/resources

Core Tools

1. search_papers — Find papers by query

curl -s "https://asta-tools.allen.ai/mcp/v1" \
  -H "Content-Type: application/json" \
  -H "x-api-key: $ASTA_API_KEY" \
  -d '{
    "jsonrpc": "2.0",
    "method": "tools/call",
    "id": 1,
    "params": {
      "name": "search_papers",
      "arguments": {
        "query": "THBS2 tumor microenvironment macrophage",
        "limit": 15
      }
    }
  }'

Returns: paper IDs, titles, authors, year, citation count, abstract snippets.

2. get_papers — Retrieve paper details by ID

Supports multiple ID types: DOI, arXiv ID, PMID, Semantic Scholar CorpusId.

curl -s "https://asta-tools.allen.ai/mcp/v1" \
  -H "Content-Type: application/json" \
  -H "x-api-key: $ASTA_API_KEY" \
  -d '{
    "jsonrpc": "2.0",
    "method": "tools/call",
    "id": 1,
    "params": {
      "name": "get_papers",
      "arguments": {
        "paper_ids": ["PMID:32273438", "DOI:10.1038/s41586-024-07487-w"],
        "fields": ["title", "authors", "year", "abstract", "citationCount", "references"]
      }
    }
  }'

3. get_citations — Citation graph traversal

curl -s "https://asta-tools.allen.ai/mcp/v1" \
  -H "Content-Type: application/json" \
  -H "x-api-key: $ASTA_API_KEY" \
  -d '{
    "jsonrpc": "2.0",
    "method": "tools/call",
    "id": 1,
    "params": {
      "name": "get_citations",
      "arguments": {
        "paper_id": "PMID:32273438",
        "direction": "citations",
        "limit": 20
      }
    }
  }'

Use "direction": "references" for backward citations.

Integration with Multi-Source Search

Add Asta as the fourth channel in the standard literature search block:

echo "=== PubMed ===" && \
curl -s "https://eutils.ncbi.nlm.nih.gov/entrez/eutils/esearch.fcgi?db=pubmed&retmode=json&retmax=20&sort=relevance&term=QUERY" && \
echo -e "\n=== OpenAlex ===" && \
curl -s "https://api.openalex.org/works?search=QUERY&per_page=10&sort=relevance_score:desc&select=id,title,authorships,publication_year,cited_by_count,doi,primary_location" && \
echo -e "\n=== Semantic Scholar ===" && \
curl -s "https://api.semanticscholar.org/graph/v1/paper/search?query=QUERY&limit=10&fields=title,authors,year,abstract,citationCount,externalIds,url" && \
echo -e "\n=== Asta (225M papers, full-text index) ===" && \
curl -s "https://asta-tools.allen.ai/mcp/v1" \
  -H "Content-Type: application/json" \
  -H "x-api-key: $ASTA_API_KEY" \
  -d '{"jsonrpc":"2.0","method":"tools/call","id":1,"params":{"name":"search_papers","arguments":{"query":"QUERY","limit":15}}}'

When Asta Adds Unique Value

Scenario	Why Asta helps
Full-text keyword search	PubMed only searches title/abstract; Asta indexes 285M+ passages from 12M full-text papers
Finding methods/protocols	Search for specific techniques mentioned only in methods sections
Citation network depth	2.4B+ citation edges enable deep forward/backward chain analysis
Cross-disciplinary discovery	225M papers across all fields, not limited to biomedical
Preprint coverage	Includes arXiv, bioRxiv, medRxiv alongside published papers

Fallback Behavior

If ASTA_API_KEY is not configured or the API returns an error:

• Skip Asta silently
• Continue with PubMed + OpenAlex + Semantic Scholar
• Do NOT report the failure to the user unless they explicitly asked for Asta results

Response Parsing

Asta MCP responses follow JSON-RPC format:

{
  "jsonrpc": "2.0",
  "id": 1,
  "result": {
    "content": [
      {
        "type": "text",
        "text": "... JSON string with paper results ..."
      }
    ]
  }
}

Parse result.content[0].text as JSON to extract paper data. Handle nested JSON strings.