Dataset Research API
Ask a question about a specialist dataset collection and get a cited answer back. These datasets are only reachable through a research run, so the API is async: start a task, then poll until the answer is ready, usually within 15 seconds to 2.5 minutes.
What is Dataset Research API?
Starts a research task scoped to one collection and returns a task id. Polling the task returns its status, then a Markdown answer with numbered citations and the list of sources once it completes.
- Endpoint
- POST /api/research
- Authentication
- API key · x-api-key header
- Cost
- 100 credits / call
- Rate limit
- 20 req / min
- Response
- Structured JSON
Try Dataset Research Live
Enter your API key and run a real query against the live endpoint.
Don't have an API key?
Sign in to your account to create and manage your API keys.
Integration guide
Copy a snippet, replace your API key, run. Works in any HTTP client — examples below in cURL, JavaScript, and Python.
/api/researchhttps://www.apipick.comStarts a research task scoped to one collection and returns a task id. Polling the task returns its status, then a Markdown answer with numbered citations and the list of sources once it completes.
querystringrequiredThe research question (max 2000 characters)
collectionstringrequiredvehicle-safety | short-sellers | sanctions
curl -X POST "https://www.apipick.com/api/research" \
-H "Content-Type: application/json" \
-H "x-api-key: YOUR_API_KEY" \
-d '{
"query": "What recalls and NHTSA investigations exist for the 2021 Tesla Model Y?",
"collection": "vehicle-safety"
}'{
"query": "What recalls and NHTSA investigations exist for the 2021 Tesla Model Y?",
"answer": "**Summary:** … [[1]](https://example.com/source)",
"results": [
{
"title": "Example source",
"url": "https://example.com/source",
"snippet": "Short excerpt the answer cites…"
}
],
"result_count": 1,
"credits_used": 100,
"remaining_credits": 99
}Rate limits
Throttling is per API key, sliding 60-second window. Hit the limit and you get a clean 429 with a Retry-After header.
20req/min
Per API key, per endpoint. Sliding 60-second window.
3concurrent
Max simultaneous in-flight requests per API key.
X-RateLimit-LimitMaximum requests allowed per minuteX-RateLimit-RemainingRequests remaining in the current windowX-RateLimit-ResetSeconds until the current window resetsRetry-AfterSeconds to wait before retrying (only on 429)HTTP/1.1 429 Too Many Requests
Retry-After: 12
X-RateLimit-Limit: 20
X-RateLimit-Remaining: 0
X-RateLimit-Reset: 12
{
"error": "rate_limit_exceeded",
"message": "Rate limit exceeded: 20 requests/minute per API key. Retry after 12s.",
"retry_after": 12
}Frequently Asked Questions
Why is this async when Cited Answer is not?
Research scoped to these collections takes longer, from 15 seconds to about 2.5 minutes. POST /api/research returns a task_id right away; poll GET /api/research/tasks/{task_id} every few seconds until status is completed.
Which collections are available?
vehicle-safety covers NHTSA recalls, owner complaints, defect investigations, and VIN decoding. short-sellers covers activist short-seller reports and long-side fund letters and pitches. sanctions covers OFAC SDN and consolidated lists, UN, UK HMT, Australia DFAT, Swiss SECO, the US Consolidated Screening List, and INTERPOL Red Notices.
When am I charged?
Only when a task completes: the first poll that returns the answer deducts 100 credits (≈ $0.10), and later polls of the same task are free. Polling a running task is free, and failed tasks cost nothing.
Can I use the sanctions collection for compliance screening?
No. It is a research aid: answers can miss listings and are not a substitute for a screening provider or a check against the official lists.
Tool schema for OpenAI / Claude?
GET /api/research/tool-schema returns ready-to-paste OpenAI function and Claude tool-use definitions, plus the list of collections.
Explore other search APIs
Real-time semantic web search built for LLM tool calling. Returns ranked titles, URLs, and clean snippets pre-shaped for agent consumption. Country and date filters supported.
Real-time news search across major outlets. Date-range and country filtering for time-sensitive queries. Built for morning briefings, market-news agents, and RAG pipelines.
Ask a focused question and get a short answer with numbered citations, researched live across the web, news, and specialist datasets. One call, no RAG pipeline to build.
Search Associated Press wire news as full article text, not snippets, over a rolling 30-day window. Built for current-events grounding and breaking-news agents.
Extract clean readable content from up to 25 URLs per call. Strips ads, nav, and boilerplate; returns markdown-flavoured text ready for LLM ingestion. 2 credits per URL.
Search peer-reviewed papers and pre-prints across arXiv, PubMed, bioRxiv, and medRxiv from one endpoint. Built for AI-driven literature review, RAG over scientific corpora, and citation extraction.
Search clinical trials, FDA drug labels, and ChEMBL bioactivity. Built for medical research, drug repurposing, and AI-driven clinical decision support workflows.
Search SEC filings (10-K, 10-Q, 8-K), US earnings call transcripts, and equity statistics. Built for AI-driven due diligence, fundamental analysis, and financial RAG pipelines.
Semantic search over global patent filings across USPTO, EPO, WIPO, and major national offices. Built for prior-art research, IP landscaping, and AI-driven competitive intelligence.
Search Polymarket and Kalshi prediction-market contracts on politics, economics, sports, and current events. Built for crowd-forecast retrieval and probability-grounded LLM answers.
Semantic search across UK case law and primary legislation from one endpoint. Built for legal research, compliance review, statutory interpretation, and AI-driven legal-tech workflows.
Search CISA Known Exploited Vulnerabilities, NVD CVE records, EPSS exploit scores, and MITRE ATT&CK techniques. Built for AI-driven vulnerability triage, threat intelligence, and security operations.
Search global and US equities, crypto, forex, ETFs, mutual funds, commodities, and US market movers. Built for AI-driven price lookups, market-data retrieval, and trading research.
Search FRED, US Bureau of Labor Statistics, World Bank indicators, IMF macro data, US federal spending, and German labour statistics. Built for AI-driven macroeconomic research and analysis.
Search US public-company balance sheets, income statements, cash flow statements, dividends, and insider transactions. Built for AI-driven fundamental analysis and due diligence.
Search the US earnings calendar for upcoming report dates, EPS/revenue estimates, and before/after-market timing. Built for AI-driven event-driven trading and research.
Search ChEMBL bioactivity, PubChem chemical structures, and Open Targets target-disease associations. Built for AI-driven drug discovery, cheminformatics, and biomedical research.
Search WHO global health statistics, NIH research grants, openFDA drug adverse-event reports, and FDA drug and device recalls. Built for AI-driven public-health research and pharmacovigilance.
Search UK Parliament debates, written questions, and member activity (Hansard). Built for AI-driven civic-tech tools, legislative monitoring, and policy research.
Search the EU MiCA crypto rulebook: the regulation, Level 2 technical standards, ESMA and EBA guidelines, and national laws in all 27 member states.
Search licensed Wiley finance journals and textbooks. Built for AI-driven academic-grade financial research, literature review, and investment analysis.