Firecrawl Research Read Paper

0 credits/call
ae_firecrawl_research_read_paper

Read the passages in one specified paper that are most relevant to your question, for example to verify a technique or benchmark result before including or rejecting the paper. Inputs: paperId (required nonempty canonical paperId or primaryId, such as arxiv:1706.03762, pmcid:PMC12530322, pmid:40953549, or doi:10.1016/j.neunet.2025.108095); question (required nonempty string); and k (optional integer 1–50, default 4) for the number of passages to return. No additional inputs are accepted. The result contains a passages array whose items may include text as a string or null, plus optional agent_hints guidance strings. If no full text is indexed, passages is empty; the service may also provide an unavailability notice. This retrieves relevant excerpts from one paper, not the complete document or a search across papers. Missing passages do not establish that a claim is false. Results are returned directly by the configured synchronous mapping, with no asynchronous task submission, status polling, or separate result-retrieval step.

Academic
0 credits/call

Pricing

0 credits/call
ConditionCharge
Each call0 credits/call

Each call to retrieve passages from a paper has a fixed charge of 0 credits.

Parameters

Optional parameters use the platform defaults when omitted.

ParameterTypeRequiredDescriptionDefault
kintegerNoNumber of passages to return (default 4).—
paperIdstringYesCanonical paperId or primaryId such as `arxiv:1706.03762`, `pmcid:PMC12530322`, `pmid:40953549`, or `doi:10.1016/j.neunet.2025.108095`.—
questionstringYesQuestion used to retrieve the most relevant full-text passages from the specified paper.—

Invocation example

Copy the request and explore its response example.

POST/agent-api/v1/tool/executeX-Api-Key
Request coderequest.sh
curl --request POST \
  --url 'https://agentearth.ai/agent-api/v1/tool/execute?tool_name=ae_firecrawl_research_read_paper' \
  --header "X-Api-Key: $AGENTEARTH_API_KEY" \
  --header 'Content-Type: application/json' \
  --data-raw '{
  "params": {
    "k": 6,
    "paperId": "pmid:38019634",
    "question": "Which exact twin support vector regression formulation is used, Peng TSVR 2010 or Khemchandani TWSVR 2016? How LSTM is trained and hidden features passed to TWSVR? chronological splits naive baseline?"
  }
}'
Response exampleresponse.json
JSON
"(no full-text passages available for this paper)"