The financial research API built for AI agents Finance Research API

A financial research API answers questions about companies, markets and the economy with a written, cited response rather than a raw data feed. It runs searches across financial sources such as filings, fundamentals, prices, transcripts and news, reads them, and synthesizes an answer with references. Developers use one to add analyst-style research to products, agents and internal tools without building retrieval, ranking and citation logic themselves.

The same shape as the Research API. content is a Markdown answer with inline citations such as [[1, 2]], or a JSON object when you send an output_schema. content_type is text or object. sources lists the financial sources read and cited, each with a URL, title and relevant excerpts, such as an SEC filing, an earnings release or an analyst report.

Company fundamentals including earnings, estimates, valuation ratios, industry-normalized financials and segment metrics. Market activity such as M&A, buybacks, capital raises, IPOs and VC funding for public and private companies. Equity prices and volume across 42 global exchanges, plus crypto, FX, metals, commodities, interest rates, inflation and real estate. Alternative signals like web traffic and app analytics. Macroeconomic indicators. And financial content: SEC filings, earnings transcripts, analyst coverage and financial news.

Both follow the same request and response pattern. The difference is the index. The Research API searches the open web and supports source_control for domain filtering. The Finance Research API searches a purpose-built financial index, accepts deep or exhaustive effort, and supports output_schema but not source_control. Use it for filings, prices, fundamentals and financial news; use the Research API for general-purpose questions.

Those are structured data feeds: you request a time series or a statement and build the analysis yourself. The Finance Research API produces the analysis: it reads, cross-references and synthesizes data into a sourced answer. It is not a time series or tick data API. If you need raw OHLCV exports or full historical series for quantitative models, use a dedicated data provider. If you need a researched, cited answer to a financial question, use this.

In an independent study on FinSearchComp T3 Global, the hardest tier of a public financial search benchmark, the API alone scored 58%, above the 51% human expert baseline and above every frontier model tested with its native web search, including Grok 4 at 51%, GPT-5 Thinking at 48%, Fable 5 at 37% and Gemini 2.5 Pro at 25%. A thin orchestration layer on top reached 64.3% with zero regressions. On You.com's own same-harness runs, the deep tier scores 87.29% on the simpler T2 tier and 58.33% on T3, ahead of Perplexity, Parallel, Tavily and Exa research endpoints. The harness is open source, so you can reproduce both results and run your own questions.

deep ($110 per 1,000 calls, under 120 seconds) is the default and handles most financial questions well: earnings summaries, sector comparisons and multi-company benchmarks. exhaustive ($500 per 1,000 calls, under 300 seconds) is for the highest quality result when response time is not a constraint, such as full 10-K analysis, deep due diligence or complex cross-market research.

Yes. Send an output_schema and content comes back as a JSON object that follows it, with content_type set to object. This is how teams extract researched figures such as margins, growth rates or guidance into typed fields for models, dashboards or downstream agents, while keeping the citations alongside.

Earnings analysis that summarizes a quarter's results, segment performance and guidance. Competitive benchmarking across companies and metrics. Due diligence on financial health, capital structure and risk factors with traceable citations. Macroeconomic research on rates, inflation and sector impact. And regulatory and filing analysis, such as what changed in a company's revenue recognition policy between annual reports.

Yes, and you should for anything that informs investment decisions, client reports or regulatory filings. Every claim links to a source such as an SEC filing, earnings release or transcript. Treat citations as starting points: follow them to the primary source and confirm the numbers before acting. The API is a research accelerator, not a substitute for primary source verification.

Under 120 seconds for deep and under 300 seconds for exhaustive. The Finance Research API has no background mode, so every request holds the connection until the answer is ready. Set your HTTP client timeout above the tier's upper bound. In the independent T3 study, median latency was 139 seconds per call.

Yes to both. The You.com MCP server exposes finance research as a tool at api.you.com/mcp/finance for Claude, Cursor, VS Code and other MCP clients. Agents can also pay per call without an account: POST /v1/finance_research accepts machine payments at the same per-tier prices, settling each request in USDC from a funded wallet instead of drawing down credits.

New accounts get $100 in free credits with no credit card required, which covers roughly 900 deep-tier calls. That is enough to evaluate answer quality on your own financial questions and build a proof of concept before paying anything.

Pricing is fixed per tier and billed per call: $110 per 1,000 calls for deep and $500 per 1,000 calls for exhaustive. Citations, sources and structured output are included. There are no token charges, minimums or platform fees, and volume discounts and annual pricing are available for high-throughput applications.

No. Your questions, answers and sources are never used for model training. Zero Data Retention (ZDR) is available for enterprise agreements and automatically purges request and response content. The platform is SOC 2 certified and DPA-ready, which matters for financial institutions with strict data handling requirements.