Documentation
Connect an agent to financial context.
Unquant is a remote financial-data MCP server. One Streamable HTTP endpoint gives compatible AI agents discoverable tools for market data, company fundamentals, processed news, US macroeconomics, and congressional transaction disclosures.
https://unquant.ai/mcp
Research guides
Practical workflows for financial-data agents
Each guide explains which public contracts to use, how to align dates, and which limitations an agent must preserve.
Start here
Three steps for a useful first result
- Connect.Add
https://unquant.ai/mcpas a remote Streamable HTTP server in a compatible MCP host. - Discover.Let the host call
tools/list, or begin withdatasets.searchwhen the required dataset is unclear. - Query carefully.Use the narrowest tool, retain dates and units, read warnings, and state freshness or delay in the answer.
Coverage
17 live data contracts
Every dataset publishes a normalized schema, supported filters, freshness, row limits, and known limitations. Raw upstream payloads are not part of the public contract.
Machine-readable references
Contracts for agents and developers
Research boundaries
Financial information, not execution.
Unquant is designed for research workflows. Market prices are delayed or end of day, public filings can be reported after the underlying event, and generated news or macro analysis can contain model error. It is not a trading venue or investment adviser.
Read the methodology and limitations