Deep async AI research agent that searches datasets and returns structured, cited results via MPP pay-per-call.
What it does
Tako's Deep Search endpoint (`/api/mpp/v1/search/deep`) is an asynchronous AI research agent that accepts a natural-language query and conducts in-depth data analysis across Tako's index of licensed, structured data sources as well as the open web. The endpoint returns HTTP 202 on acceptance, and results are retrieved by polling the companion `/api/mpp/v1/search/deep/status` endpoint (which is free, authenticated by the original payment receipt). Results include structured knowledge cards with charts, tables, source citations, and methodology notes.
The endpoint is part of Tako's broader MPP-enabled API suite, which also includes fast search ($0.037/call), data visualization ($0.037/call), chart editing ($0.007/call), ThinViz card creation ($0.004/call), deep SSE-streamed search ($0.55/call), and full research report generation ($5.50/call). Deep Search is priced at approximately $0.55 per request, settled via the Tempo method on chain 4217 using pathUSD (6 decimals; the raw challenge amount is 549999 base units). No API key or signup is required — payment is handled entirely through the MPP 402 challenge-response flow.
Tako emphasizes data accuracy and credibility: each result cites its sources, includes timestamps, and where applicable provides methodology notes explaining how data was calculated. The platform is SOC 2 Type II compliant. Request bodies are JSON POSTs, though the exact schema for the query payload is not fully documented in the OpenAPI spec (the schema is listed as a generic object). Developers should consult Tako's full documentation at docs.tako.com for field-level details.
Capabilities
Use cases
- —Company and investment analysis — evaluate performance, fundraising, M&A, and benchmarking
- —Market and macroeconomic research with cited, structured data
- —Embedding authoritative data answers into LLM agent workflows
- —Generating visual knowledge cards (charts, tables, maps) from natural-language questions
- —Competitive intelligence and strategy research backed by licensed data sources
Fit
Best for
- —AI agents that need cited, structured data answers without managing API keys
- —Research and strategy teams needing deep, evidence-backed analysis
- —Developers embedding data-grounded AI responses into products
Not for
- —Real-time sub-second lookups (this is async with polling; use /search/fast instead)
- —Free or high-volume bulk data scraping (each call costs ~$0.55)
- —Querying private datasets (use Tako Connect for that)
Quick start
curl -X POST https://tako.com/api/mpp/v1/search/deep \
-H "Content-Type: application/json" \
-d '{"query": "How has OpenAI revenue grown since 2022?"}'
# Returns 402 with WWW-Authenticate: Payment header.
# Complete the MPP Tempo payment, then re-send with receipt.
# Poll /api/mpp/v1/search/deep/status?id=<task_id> for results.Example
Request
{
"query": "How has OpenAI revenue grown since 2022?"
}Endpoint
Quality
The endpoint is live (402 MPP challenge confirmed), pricing is clear from both the OpenAPI spec and the decoded challenge, and the platform's purpose is well-documented on the landing page. However, the OpenAPI request/response schemas are generic objects with no field-level detail, and no example response is available. The exact request body fields must be inferred.
Warnings
- —Request body schema is a generic object — exact required fields are undocumented in the OpenAPI spec
- —No example response payload is available; response structure must be discovered via docs.tako.com
- —Currency address 0x20c000000000000000000000b9537d11c60e8b50 assumed to be pathUSD with 6 decimals based on Tempo/chain 4217 context; if incorrect, the $0.55 price estimate may be wrong
- —robots.txt blocks major AI crawlers (GPTBot, ClaudeBot, CCBot, etc.) from tako.com
Citations
- —Deep Search is priced at 549999 base units via Tempo on chain 4217https://tako.com/api/mpp/v1/search/deep
- —Tako is SOC 2 Type II complianthttps://tako.com/pricing
- —Tako knowledge cards include clear sources, timestamps, and methodology noteshttps://tako.com
- —Fast Search and Visualize are priced at $0.036667 per call; reports at $5.50https://tako.com/api/mpp/v1/search/deep
- —Full developer documentation is at docs.tako.comhttps://tako.com
- —Polling endpoint /api/mpp/v1/search/deep/status is free and receipt-authenticatedhttps://tako.com/api/mpp/v1/search/deep