Perplexity launched Finance Search in its Agent API on May 6, 2026 — a single tool call that returns licensed financial datasets, real-time market data, and cited web sources for AI agents that need current, verifiable financial answers. Perplexity Finance Search bundles licensed data feeds (prices, fundamentals, transcripts, estimates, filings, ETF holdings, market activity) with retrieval against the broader web, eliminating the need for developers to integrate each licensed financial data provider separately. Pricing is $5 per 1,000 invocations, billed separately from model tokens. The launch positions Perplexity’s Agent API as a competitive option for builders of financial AI applications who previously had to assemble data feeds from Bloomberg, Refinitiv, FactSet, and other providers — each with its own contract, integration, and licensing complexity.
What’s actually new
Finance Search consolidates several distinct capabilities into a single tool call accessible from the Perplexity Agent API. The capabilities span structured market data (real-time prices, historical data, technical indicators, options data), fundamental data (financial statements, ratios, estimates, ownership), corporate actions and events (earnings, dividends, splits, M&A), and unstructured financial content (filings, earnings call transcripts, analyst reports, financial news with citations). The single-tool-call abstraction is the key product innovation — agents request what they need in plain language and get structured, citable results.
The pricing structure is competitive. $5 per 1,000 invocations sits well below the licensed-data-provider equivalents that financial-services customers typically pay. Bloomberg Terminal subscriptions run thousands of dollars per user per month; Refinitiv and FactSet contracts run hundreds of thousands to millions for enterprise access; specialized data feeds for retail-facing applications carry usage-based licensing that can quickly become substantial. Perplexity’s pricing democratizes access for smaller developers and consumer-facing applications that couldn’t justify the enterprise data licensing costs.
The Agent API surface itself has matured. Perplexity’s API extends beyond search to a structured agent-tool-use model where developers compose agents from primitives like Finance Search, Web Search, structured retrieval, and standard model inference. Finance Search joins the catalog as a domain-specific tool. The API’s documentation, SDKs, and developer experience are competitive with the major foundation-model APIs in 2026.
The licensing arrangements behind Finance Search matter. Perplexity has acquired or licensed access to multiple licensed financial data providers; Finance Search abstracts the licensing complexity from developers. Developers using Finance Search are licensed to use the data through Perplexity’s terms; they don’t need separate contracts with the underlying data providers. This is a meaningful abstraction — it lets developers ship financial AI applications without the legal and procurement work that historically delayed launches.
The cited-source pattern is essential for financial use cases. Every Finance Search result includes citations to the underlying data sources, which lets agents produce auditable outputs. The citation pattern has been a Perplexity hallmark since the consumer product launched; bringing it to the API makes it accessible to developers building applications where source attribution matters — investor-facing tools, regulated financial communications, and similar contexts.
Why it matters
- Financial AI just got cheaper to build. The historical cost barrier of licensing financial data feeds individually has shut out smaller developers and consumer-facing applications. $5 per 1K calls is accessible enough that hobbyist developers, small startups, and mid-market companies can build financial AI applications without enterprise-scale data budgets.
- The financial AI app surface expands. When a single API call replaces months of vendor-specific integration work, applications that wouldn’t have been viable become economic. Expect a wave of new financial AI products from indie developers and small teams through Q3-Q4 2026.
- Bloomberg, Refinitiv, and FactSet face pressure. The traditional financial data oligopoly has been resilient because their licensing terms shut out competitors. Perplexity’s aggregation undermines that moat. Whether the incumbents respond with their own AI-friendly APIs or maintain the high-friction enterprise licensing model is the strategic question.
- Anthropic‘s financial agents get a complement. Anthropic’s 10 financial-services agents launched May 5 need data feeds; Finance Search is a natural complement. Builders combining Anthropic agents with Perplexity Finance Search produce capable financial-AI workflows faster than either alone.
- The cited-source pattern moves further into financial workflows. Citations matter in financial contexts both for compliance (regulatory expectations on source attribution) and for trust (financial advice or analysis without sources is increasingly unacceptable). Finance Search’s citation-by-default reinforces this norm.
- Perplexity’s strategic position evolves. The company has been positioned primarily as a consumer search alternative to Google. The Agent API and Finance Search position Perplexity as serious infrastructure for AI builders, particularly in financial applications. The broader Agent API strategy makes Perplexity a meaningful competitor to OpenAI’s API, Anthropic’s API, and Google’s Gemini API in specific domains.
How to use Perplexity Finance Search today
Three steps put a developer on Finance Search.
- Get an API key. Sign up at perplexity.ai/api and provision an API key. The Agent API is generally available; no waitlist as of mid-May 2026. Add billing for production use.
- Read the Agent API docs. The documentation at docs.perplexity.ai includes the Finance Search tool’s parameters, response formats, and example usage patterns. Specific attention to the citation format and licensing terms is worth the time before substantial integration.
- Integrate the tool into your agent. The Agent API uses standard tool-use patterns; Finance Search is one tool among others. Compose with model inference, web search, and other tools as needed for your application.
A reference integration in Python:
MASK12
For agentic workflows that combine Finance Search with model reasoning, the pattern uses Finance Search as a tool the model can invoke:
MASK13
The cost calculation for production applications. At $5 per 1K Finance Search invocations plus model token cost, a typical financial-research agent making 10 Finance Search calls per session costs about $0.05 per session for data plus the model inference cost. For applications running thousands of sessions daily, the monthly Finance Search cost is meaningful but well below the cost of equivalent direct data feed licensing.
How it compares
The financial data API landscape in 2026 has multiple options for AI applications. The table below summarizes the leaders along the dimensions that matter for AI agent integration.
| Provider | Coverage | API style | Pricing | Best fit |
|---|---|---|---|---|
| Perplexity Finance Search | Prices, fundamentals, filings, estimates, market | Single agent tool call | $5 / 1K invocations | AI agents, smaller developers |
| Bloomberg Terminal API (BLPAPI) | Comprehensive financial data | SDK / proprietary | $24K-30K per terminal/year | Enterprise financial users |
| Refinitiv (LSEG) Eikon Data API | Comprehensive financial data | REST + SDK | Enterprise contracts | Buy-side and sell-side institutions |
| FactSet | Comprehensive financial data | SDK / FQL queries | Enterprise contracts | Asset managers, IBs |
| Polygon.io | Market data, US equities focus | REST + WebSocket | $29-2K/month tiers | Mid-market developers |
| Alpha Vantage | Limited market data | REST | Free + paid tiers | Hobbyist developers |
| Financial Modeling Prep | Fundamentals + market | REST | $15-99/month tiers | Indie developers |
| Tiingo | Market data + fundamentals | REST + WebSocket | $10-499/month tiers | Quant individual / small teams |
Two takeaways. First, Perplexity Finance Search occupies a distinct niche — the cited, agent-friendly aggregation pattern that other providers don’t offer in this format. The traditional providers (Bloomberg, Refinitiv, FactSet) deliver more comprehensive raw data through enterprise contracts; the developer-friendly providers (Polygon, Tiingo, FMP) deliver structured access at moderate cost; Perplexity adds the agent-API aggregation that simplifies integration for AI applications. Second, the right choice depends on the use case. High-frequency trading or institutional research need the depth of Bloomberg/Refinitiv. Consumer-facing apps and AI agent applications increasingly favor the Perplexity-style aggregation. Multi-source architectures combining several providers are common at scale.
What’s next
Three things to watch over the next two quarters. First, additional domain-specific tools in the Perplexity Agent API. Finance Search is the first deeply-domain-specific tool; expect tools for legal research, healthcare data, real estate, and other domains through 2026. The pattern of “single tool call for a domain’s licensed data” is replicable. Second, competitive responses from Bloomberg, Refinitiv, and FactSet. The incumbents’ enterprise licensing model has been resilient but Perplexity’s pressure could force AI-friendly API options. Third, the regulatory dimension. Financial data licensing has evolved over decades with specific terms about redistribution and use; the Perplexity aggregation pattern raises licensing questions that will likely be tested in coming quarters.
The longer-term implication is that the financial AI application surface is broadening rapidly. Applications that combine financial data, AI reasoning, and citation-rigor produce capabilities that consumer financial apps haven’t matched. The 2027 cohort of financial AI startups will likely produce several breakout consumer and prosumer applications enabled partly by infrastructure like Finance Search. Watch for new entrants in personal finance, investment research, and small-business financial management.
Frequently Asked Questions
Is Perplexity Finance Search compliant for use in regulated financial contexts?
The licensed data flows through Perplexity’s commercial agreements and is licensed for use through your Perplexity contract. For specific regulated use cases (registered investment advisors, broker-dealers, banks providing customer-facing financial information), validate with your compliance team that Perplexity’s terms align with your regulatory obligations. The cited-source pattern helps with audit trail requirements but doesn’t substitute for compliance review.
How current is the data Finance Search returns?
Real-time market data is delivered with low latency from licensed feeds. Fundamental data is updated as filings are processed (typically same-day for major filings). News and qualitative content is updated continuously. Specific staleness depends on the data category; the documentation specifies update cadence per data type.
Can I use Finance Search results in customer-facing applications?
Yes, with appropriate disclosures about data sources and the AI-driven nature of the analysis. The cited sources should be visible to end users where applicable. Perplexity’s terms specify what’s permitted; review carefully for high-stakes customer-facing use cases.
How does this compare to building my own RAG over financial filings?
Finance Search aggregates licensed data feeds (prices, structured fundamentals, real-time data) plus retrieval. Building your own RAG over filings handles the unstructured content but doesn’t access licensed data. For comprehensive financial AI, Finance Search plus custom RAG over your specific document corpus is often the right combination.
Will Perplexity Finance Search work with non-Perplexity LLMs?
Finance Search is part of the Perplexity Agent API and is most easily used with Perplexity’s Sonar models. Developers can also call Finance Search directly and pass the results as context to other LLMs (Claude, GPT, Gemini) — the data and citations work regardless of which model produces the final analysis. The cleanest integration is end-to-end on Perplexity, but the data is portable.
What about international markets and non-US equities?
Coverage spans major global exchanges with varying depth by market. US, UK, EU, and major Asian markets have the deepest coverage. Specific market coverage is detailed in the documentation. For applications targeting specific non-US markets, validate coverage before committing to integration.