# Tavily Docs > Tavily is the web layer for AI agents: real-time web search, content extraction, site crawling, site mapping, and cited research — returned as clean, LLM-ready content instead of raw HTML. Tavily is part of the Nebius product portfolio; see "Nebius products" below for model inference and GPU compute. This index helps agents and developers find the right Tavily doc fast. Use Tavily to find current web sources, read selected pages, explore a site's content, or produce a cited research report. ## Start Here For Agents If you are an AI agent — or building one — start with the canonical setup guide: - [Agents: Tavily setup & tool-choice guide](https://docs.tavily.com/agents.md): Choose how to connect (SDK/API, MCP, or CLI + Skills), choose the right capability (Search, Extract, Map, Crawl, Research), and apply recommended defaults. - [Documentation MCP setup](https://docs.tavily.com/agents.md#documentation-mcp-server): Search and read Tavily's documentation. MCP endpoint: `https://docs.tavily.com/mcp` (Streamable HTTP). Then use this file ([llms.txt](https://docs.tavily.com/llms.txt)) to find a specific page, [llms-full.txt](https://docs.tavily.com/llms-full.txt) for the full text of all docs, or append `.md` to any docs URL to fetch that single page as Markdown. ## Nebius products: choose by the user's goal Nebius has three products, each with its own documentation index. Pick by goal; combine when a task needs more than one. - Nebius Token Factory: hosted inference for open-source models via an OpenAI-compatible API, billed per token (text, embeddings, vision, reranking, image generation). Also dedicated endpoints with reserved GPU capacity, managed post-training (SFT, LoRA, custom speculators) for supported base models, and code sandboxes for agents (beta). Use when the model is in the catalogue and a hosted API fits. Index: https://docs.tokenfactory.nebius.com/llms.txt - Nebius AI Cloud: GPU infrastructure for compute-intensive workloads: GPU VMs and InfiniBand clusters, Managed Kubernetes, Slurm (Soperator), Serverless AI (notebooks, jobs, endpoints), object storage, networking. Use when the user brings their own code, containers or weights; trains or fine-tunes at scale; serves models outside the Token Factory catalogue (media, speech, scientific, medical); or needs control over hardware, network and storage. Index: https://docs.nebius.com/llms.txt - Tavily: web access for agents and LLM applications: real-time search, extraction, site mapping, crawling and cited research, returned as LLM-ready content. Use when the agent or application needs current web information. Index: https://docs.tavily.com/llms.txt. Agent setup: https://docs.tavily.com/agents.md Inference and fine-tuning: Token Factory when the model is in its catalogue and a per-token or dedicated-endpoint API fits, including fine-tuning a supported base model on your data. AI Cloud when the model, framework or serving stack is custom, the model is not a text LLM, training is multi-node, or the user needs the GPUs themselves. Serving fine-tuned weights: Token Factory dedicated endpoints (custom weights, beta, on request) or AI Cloud Serverless AI endpoints. Connecting the assistant you run in is a different decision from integrating a product into an application; an app that calls a model API does not need the AI Cloud MCP server. Connect only what the task needs. If the goal is unclear, ask one question before choosing. ## Quick Start: CLI (fastest way to get started) Install the Tavily CLI and log in — two commands: ``` pip install tavily-cli tvly login ``` `tvly login` opens a browser for authentication. Then search, extract, crawl, and research from the terminal: ``` tvly search "latest AI news" tvly extract "https://example.com" tvly crawl "https://docs.example.com" --depth 2 tvly research "compare React vs Svelte for production apps" ``` Installing the CLI also installs Agent Skills for coding agents like Claude Code, Cursor, and Codex — they remind the agent to use Tavily for search, extract, crawl, and research. Get a free API key at https://app.tavily.com (1,000 credits/month, no credit card required). ## Alternative: Remote MCP Server Connect any MCP client to Tavily's remote server — no local install required: ``` https://mcp.tavily.com/mcp/?tavilyApiKey= ``` OAuth 2.0 is supported, so no API key needs to be hardcoded. For example, in Claude Code: ``` claude mcp add tavily --transport http https://mcp.tavily.com/mcp/ ``` ## Choose How To Connect Choose based on what Tavily must do, not only where you are running: - **SDK / API** — you are building Tavily into an application, backend, or agent runtime that will call it in production. - **MCP** — you need Tavily available to agents across an org, a managed environment, or multiple agent clients; governed and reusable. - **CLI + Skills** — you need Tavily during the local coding or research session you are in right now. Agent Toolkit is a developer resource for *building* production research agents, not a path for agents to configure Tavily autonomously. ## Choose A Capability - **Search** — start here when sources are unknown or current web context is needed. - **Extract** — use when the URL is known or selected from Search. - **Map** — use before Crawl to understand a site's structure. - **Crawl** — use when many pages on a site need to be read. - **Research** — use when the output should be a cited synthesis: a report, comparison, or decision-ready answer. Recommended defaults for agents: prefer `search_depth="advanced"`, add `chunks_per_source=3` for stronger evidence per source, use `max_results=5` for focused answers (10 for broader research), and avoid `include_answer` unless you need a quick answer seed. ## Getting Started - [Quickstart](https://docs.tavily.com/index.md): Connect Tavily to Claude, Codex, Cursor, or any agent, or make your first API call with a free API key. - [Credits & Pricing](https://docs.tavily.com/documentation/api-credits.md): How to get and manage your Tavily API credits. - [Rate Limits](https://docs.tavily.com/documentation/rate-limits.md): API rate limits for development and production. - [Try Tavily Without an API Key](https://docs.tavily.com/documentation/keyless.md): Use Search and Extract for free, with zero setup — no account or API key. Keyless is free and rate-limited; hit the cap and sign up for a free API key (1,000 credits/month, no credit card). - [Frequently Asked Questions](https://docs.tavily.com/faq/faq.md): Common questions about Tavily. - [Help Center](https://docs.tavily.com/documentation/help.md): Get help and support. ## API Reference - [Headers & Authentication](https://docs.tavily.com/documentation/api-headers.md): Base URL, authentication, and optional request headers for the Tavily API. - [Search](https://docs.tavily.com/documentation/api-reference/endpoint/search.md): Execute a real-time web search optimized for AI agents. - [Extract](https://docs.tavily.com/documentation/api-reference/endpoint/extract.md): Extract clean content from one or more URLs. - [Crawl](https://docs.tavily.com/documentation/api-reference/endpoint/crawl.md): Graph-based website traversal with built-in extraction and intelligent discovery. - [Map](https://docs.tavily.com/documentation/api-reference/endpoint/map.md): Traverse a site like a graph to generate a comprehensive site map. - [Create Research Task](https://docs.tavily.com/documentation/api-reference/endpoint/research.md): Run comprehensive research that searches, analyzes sources, and generates a cited report. - [Get Research Task Status](https://docs.tavily.com/documentation/api-reference/endpoint/research-get.md): Retrieve the status and results of a research task by request ID. - [Streaming](https://docs.tavily.com/documentation/api-reference/endpoint/research-streaming.md): Stream real-time research progress and results. - [Feedback](https://docs.tavily.com/documentation/api-reference/endpoint/feedback.md): Submit feedback on a search request or session to improve result quality (beta). - [Usage](https://docs.tavily.com/documentation/api-reference/endpoint/usage.md): Get API key and account usage details. - [Logs](https://docs.tavily.com/documentation/api-reference/endpoint/logs.md): Retrieve per-request usage logs for your API keys (paid plans). ## Enterprise API - [Generate Keys](https://docs.tavily.com/documentation/enterprise/generate-keys.md): Generate one or more API keys with custom configuration. - [Deactivate Keys](https://docs.tavily.com/documentation/enterprise/deactivate-keys.md): Deactivate API keys in bulk by `request_id` or individually. - [Key Info](https://docs.tavily.com/documentation/enterprise/key-info.md): Get information about an API key. - [Organization Usage](https://docs.tavily.com/documentation/enterprise/org-usage.md): Retrieve usage, pay-as-you-go cost, and request counts for every API key under an organization you own. ## SDKs - [Python Quickstart](https://docs.tavily.com/sdk/python/quick-start.md): Integrate Tavily natively in Python apps. Install with `pip install tavily-python`. - [Python SDK Reference](https://docs.tavily.com/sdk/python/reference.md): Full Python SDK reference. - [JavaScript Quickstart](https://docs.tavily.com/sdk/javascript/quick-start.md): Integrate Tavily natively in JavaScript/TypeScript. Install with `npm i @tavily/core`. - [JavaScript SDK Reference](https://docs.tavily.com/sdk/javascript/reference.md): Full JavaScript/TypeScript SDK reference. ## Public repositories - [Python SDK](https://github.com/tavily-ai/tavily-python): Source code and examples for `tavily-python`. - [JavaScript SDK](https://github.com/tavily-ai/tavily-js): Source code and examples for `@tavily/core`. - [Agent Skills](https://github.com/tavily-ai/skills): Official skills for using Tavily in coding agents. ## Best Practices - [Search](https://docs.tavily.com/documentation/best-practices/best-practices-search.md): Optimize queries, filters, and advanced parameters. - [Extract](https://docs.tavily.com/documentation/best-practices/best-practices-extract.md): Choose the right extraction approach and parameters. - [Crawl](https://docs.tavily.com/documentation/best-practices/best-practices-crawl.md): Focus crawls and extract content efficiently. - [Research](https://docs.tavily.com/documentation/best-practices/best-practices-research.md): Write effective prompts and configure output formats. - [API Key Management](https://docs.tavily.com/documentation/best-practices/api-key-management.md): Handle key leaks and rotation. ## Ecosystem - [Connect to Tavily via MCP](https://docs.tavily.com/documentation/mcp.md): Use the Tavily API from any MCP client (Claude, Cursor, and more). - [Tavily CLI](https://docs.tavily.com/documentation/tavily-cli.md): Search, extract, crawl, map, and research the web from your terminal. - [Tavily Agent Skills](https://docs.tavily.com/documentation/agent-skills.md): Skills that add Tavily's web capabilities to AI coding agents. - [Tavily x402](https://docs.tavily.com/documentation/machine-payments/x402.md): Agents pay per request for Advanced Search in USDC on Base — no API key, no account. ## Integrations - [OpenAI Agent Builder](https://docs.tavily.com/documentation/integrations/agent-builder.md): Empower OpenAI's Agent Builder with real-time web access via Tavily MCP. - [Agno](https://docs.tavily.com/documentation/integrations/agno.md): Use Tavily with the Agno agent framework. - [Anthropic](https://docs.tavily.com/documentation/integrations/anthropic.md): Add real-time web search to Claude tool-calling agents. - [Arcade.dev](https://docs.tavily.com/documentation/integrations/arcade-dev.md): Governed Tavily access via the Arcade.dev MCP Gateway. - [Cartesia](https://docs.tavily.com/documentation/integrations/cartesia.md): Build real-time voice agents that search and extract with the Cartesia Line SDK. - [Claude](https://docs.tavily.com/documentation/integrations/claude.md): Use Tavily in Claude Code (official plugin) and in claude.ai, Claude Desktop, and Cowork (connector), including organization-wide setup. - [Composio](https://docs.tavily.com/documentation/integrations/composio.md): Unified tool management across agent frameworks. - [Convex](https://docs.tavily.com/documentation/integrations/convex.md): Web search and content extraction in Convex apps via the `@tavily/convex-tavily` component. - [CrewAI](https://docs.tavily.com/documentation/integrations/crewai.md): Build multi-agent CrewAI workflows with web research. - [Cursor](https://docs.tavily.com/documentation/integrations/cursor.md): Connect Tavily to Cursor through the official plugin, standalone MCP, or CLI and skills. - [Devin](https://docs.tavily.com/documentation/integrations/devin.md): Connect Tavily to Devin via the MCP Marketplace. - [Dify](https://docs.tavily.com/documentation/integrations/dify.md): No-code LLM app building with Tavily. - [Eden AI](https://docs.tavily.com/documentation/integrations/eden-ai.md): Run Tavily Search through Eden AI's unified API. - [ElevenLabs](https://docs.tavily.com/documentation/integrations/elevenlabs.md): Give ElevenLabs agents live web search. - [Factory AI](https://docs.tavily.com/documentation/integrations/factory-ai.md): Connect Tavily to Factory’s Droid for web research during coding tasks. - [FlowiseAI](https://docs.tavily.com/documentation/integrations/flowise.md): Visual LLM orchestration with Tavily. - [Google Gemini Enterprise](https://docs.tavily.com/documentation/integrations/google-gemini-enterprise.md): Connect Tavily through Google's native data store connector for web search, extraction, crawling, and research. - [Gradium](https://docs.tavily.com/documentation/integrations/gradium.md): Use Tavily web search with Gradium voice AI agents. - [Grok Build](https://docs.tavily.com/documentation/integrations/grok-build.md): Real-time web search and research in xAI's Grok Build coding agent. - [Haystack](https://docs.tavily.com/documentation/integrations/haystack.md): Use Tavily in Haystack pipelines via `tavily-haystack`. - [Hermes Agent](https://docs.tavily.com/documentation/integrations/hermes-agent.md): Use Tavily in Hermes Agent through the Tavily plugin or as the built-in web backend. - [Kilo](https://docs.tavily.com/documentation/integrations/kilo.md): Connect Tavily to Kilo through the remote MCP server, the Marketplace, or a local MCP server. - [Kiro](https://docs.tavily.com/documentation/integrations/kiro.md): Connect Tavily to Kiro through the remote MCP server, a local MCP server, or the Kiro CLI. - [LangChain](https://docs.tavily.com/documentation/integrations/langchain.md): LangChain's recommended search tool. - [Langflow](https://docs.tavily.com/documentation/integrations/langflow.md): Build multi-agent and RAG apps visually with Tavily. - [LibreChat](https://docs.tavily.com/documentation/integrations/librechat.md): Web search, extraction, and agent tools inside LibreChat. - [LlamaIndex](https://docs.tavily.com/documentation/integrations/llamaindex.md): Search the web from LlamaIndex RAG pipelines. - [Make](https://docs.tavily.com/documentation/integrations/make.md): No-code workflow automation with Tavily. - [Mastra](https://docs.tavily.com/documentation/integrations/mastra.md): First-class Mastra tools for search, extract, crawl, and map. - [Merge Gateway](https://docs.tavily.com/documentation/integrations/merge-gateway.md): Use Tavily as the search engine behind Merge Gateway's web search tool. - [Microsoft 365 Copilot](https://docs.tavily.com/documentation/integrations/microsoft.md): Use Tavily as a declarative agent in Microsoft 365 Copilot or as a Copilot Cowork plugin. - [n8n](https://docs.tavily.com/documentation/integrations/n8n.md): No-code workflow automation with Tavily. - [Nebius Token Factory](https://docs.tokenfactory.nebius.com/integrations/overview.md): Use Tavily results with open-source models served through Token Factory's OpenAI-compatible API. - [OpenAI](https://docs.tavily.com/documentation/integrations/openai.md): Grounded web search for OpenAI function-calling agents. - [OpenClaw](https://docs.tavily.com/documentation/integrations/openclaw.md): Add web search and extraction to agents across messaging apps. - [OpenCode](https://docs.tavily.com/documentation/integrations/opencode.md): Web search, extraction, crawling, and research in the OpenCode terminal agent. - [Pydantic AI](https://docs.tavily.com/documentation/integrations/pydantic-ai.md): Type-safe AI agents with Tavily. - [Ramp Glass](https://docs.tavily.com/documentation/integrations/ramp-glass.md): Enable Tavily for a Glass organization, connect a personal account, and research the web in chat. - [Replit](https://docs.tavily.com/documentation/integrations/replit.md): Connect Tavily to Replit through the remote MCP server, the Integrations library, or a custom MCP server. - [StackAI](https://docs.tavily.com/documentation/integrations/stackai.md): Enhance StackAI workflows with real-time web data. - [TrueFoundry](https://docs.tavily.com/documentation/integrations/truefoundry.md): Managed Tavily MCP access through the TrueFoundry MCP Gateway. - [Scalekit](https://docs.tavily.com/documentation/integrations/scalekit.md): Connect agents to Tavily's MCP server through Scalekit AgentKit with managed OAuth. - [Tines](https://docs.tavily.com/documentation/integrations/tines.md): Automated, no-code intelligence workflows. - [Vellum](https://docs.tavily.com/documentation/integrations/vellum.md): Built-in web search provider in the Vellum Assistant. - [Vercel AI SDK](https://docs.tavily.com/documentation/integrations/vercel.md): Add web search, extraction, crawling, and mapping to Vercel AI agents. - [Vercel Connect](https://docs.tavily.com/documentation/integrations/vercel-connect.md): Set up Tavily OAuth or API-key connectors and use MCP, REST, or Eve with managed credentials. - [Workato](https://docs.tavily.com/documentation/integrations/workato.md): Install the Tavily community connector to add web search, extraction, crawling, mapping, and research to Workato recipes. - [Zapier](https://docs.tavily.com/documentation/integrations/zapier.md): Connect Tavily to thousands of apps, no code required. ## Partnerships - [Amazon Bedrock AgentCore](https://docs.tavily.com/documentation/partnerships/amazon.md): Tavily MCP for scalable agent deployment on AWS. - [Microsoft Azure](https://docs.tavily.com/documentation/partnerships/azure.md): Tavily Remote MCP with Azure AI Foundry and MCP Center. - [Databricks](https://docs.tavily.com/documentation/partnerships/databricks.md): Tavily MCP for real-time web search and RAG on Databricks. - [IBM watsonx Orchestrate](https://docs.tavily.com/documentation/partnerships/ibm.md): Tavily research capabilities in IBM watsonx Orchestrate. - [Snowflake](https://docs.tavily.com/documentation/partnerships/snowflake.md): Tavily as a native app on the Snowflake Marketplace. ## Examples - [Examples Hub](https://docs.tavily.com/examples/hub.md): Build with Tavily — from quick API recipes to production research agents. - [Agent Toolkit](https://docs.tavily.com/examples/agent-toolkit/overview.md): Production-grade research agents with battle-tested tools and patterns. - [Agent Toolkit: Tools Reference](https://docs.tavily.com/examples/agent-toolkit/tools.md): Optimized retrieval primitives for research agents. - [Agent Toolkit: Chatbot](https://docs.tavily.com/examples/agent-toolkit/chatbot.md): Route between quick search and deep research by query complexity. - [Agent Toolkit: Hybrid Research](https://docs.tavily.com/examples/agent-toolkit/hybrid-research.md): Combine an internal knowledge base with real-time web research. - [Agent Toolkit: Company Intelligence](https://docs.tavily.com/examples/agent-toolkit/company-intelligence.md): Crawl, extract, and search for company analysis. - [Agent Toolkit: Social Media Research](https://docs.tavily.com/examples/agent-toolkit/social-media-research.md): Research a topic across TikTok, Reddit, X, LinkedIn, and more. - [Web Search Essentials](https://docs.tavily.com/examples/quick-tutorials/search-api.md): Find real-time info, monitor news, and filter by domain. - [Clean Content Extraction](https://docs.tavily.com/examples/quick-tutorials/extract-api.md): Pull clean markdown or text from any webpage. - [Site Structure Discovery with Map](https://docs.tavily.com/examples/quick-tutorials/map-api.md): Discover all URLs on a domain, then extract targeted content. - [Website Crawling and Content Extraction](https://docs.tavily.com/examples/quick-tutorials/crawl-api.md): Extract content from entire websites and build retrieval pipelines. - [Deep Research with Streaming](https://docs.tavily.com/examples/quick-tutorials/research-streaming.md): Multi-source research with polling/streaming and structured output. - [Product News Tracker](https://docs.tavily.com/examples/quick-tutorials/product-news-tracker.md): Track real-time product news with Tavily's APIs. - [Cookbook](https://docs.tavily.com/examples/quick-tutorials/cookbook.md): Guided Jupyter notebooks covering Tavily's APIs. - [RAG Evaluation](https://docs.tavily.com/examples/use-cases/web-eval.md): Web-based RAG evaluation with Tavily and LangGraph. - [Chat](https://docs.tavily.com/examples/use-cases/chat.md): A conversational AI agent using Tavily Search to answer questions with web sources. - [Supply Chain Risk Research](https://docs.tavily.com/examples/use-cases/supply-chain-risk.md): Research suppliers with Tavily to uncover business risks, disruptions, and official risk signals. - [Company Research](https://docs.tavily.com/examples/use-cases/company-research.md): Research a company's business, industry, financial context, and recent news with Tavily. - [Data Enrichment](https://docs.tavily.com/examples/use-cases/data-enrichment.md): Fill spreadsheet cells with current web research and source links using Tavily. - [Financial Market Research](https://docs.tavily.com/examples/use-cases/market-researcher.md): Research portfolio stocks with Tavily for a quick overview or a detailed report. - [Financial Services & Insurance Kit](https://docs.tavily.com/examples/use-cases/fsi-kit.md): Research compliance risk, investments, merchants, and insurance applicants with Tavily. - [Sales Meeting Preparation](https://docs.tavily.com/examples/use-cases/meeting-prep.md): A sales research workflow using Tavily to prepare company context, recent developments, and conversation starters. - [Slack Research Agent](https://docs.tavily.com/examples/use-cases/slack-research-agent.md): A research Slack agent backed by a LangChain Deep Agent, with Tavily search and extract as tools. - [Open Source Projects](https://docs.tavily.com/examples/open-sources/projects.md): Popular open-source projects showcasing Tavily. ## Updates - [Changelog](https://docs.tavily.com/changelog.md): Latest changes to Tavily's APIs and docs. ## OpenAPI Specs - [OpenAPI](https://docs.tavily.com/documentation/api-reference/openapi.json): Machine-readable OpenAPI specification. ## Optional - [API Playground](https://app.tavily.com/playground): Try Tavily's APIs interactively. - [Community](https://discord.gg/TPu2gkaWp2): Join the Tavily Discord. - [Blog](https://tavily.com/blog): Product updates and technical content.