When do my monthly API credits reset?

Your credits reset on the first day of each month, regardless of when you signed up or your billing date.

What happens if I run out of credits?

If you exhaust your monthly credits, requests will stop until credits reset or you upgrade to a paid plan for higher limits.

How does Tavily handle PII and security in requests?

Tavily has built-in validation layers that can detect and block requests containing personally identifiable information (PII) like email addresses, preventing data leakage.

Can I control the depth of a web crawl or extraction?

Yes. The crawl and extract endpoints offer parameters like `extract_depth` (basic/advanced) and `max_depth` to control how thoroughly content is gathered and processed.

Is there a way to test the API before committing?

Yes. You can sign up for a free account to get 1,000 monthly credits. There's also an interactive API Playground available for testing endpoints without code.

Directory/Search/AI Assistants/Tavily

Tavily

Tavily is a production-grade API platform that provides AI agents and applications with real-time, structured access to the web. Its core capabilities include fast and relevant search, content extraction for LLMs, deep research agents, full-site crawling, and website mapping. The service is built to ground models in fresh web context, preventing hallucinations by retrieving and structuring live data. It handles billions of pages with high uptime and low latency, targeting developers and enterprises building AI-driven products that require reliable web information.

Visit website

Added on

Sep 22, 2026

Web Search APIContent ExtractionWebsite Crawl/MapResearch AgentPython/JS SDKsPII Protection

Product information

Everything worth knowing about Tavily

The complete picture — from everyday features to the technical detail builders and IT folks dig for.

Core features

What it actually does

Real-time Web Search

Fast and relevant web results with intelligent caching and indexing for low latency.

Content Extraction

Clean and structured content extraction from webpages for LLM consumption.

Website Crawling & Mapping

Extract content from entire websites and discover URLs across any domain.

Deep Research Agent

An agent for in-depth web research with structured output schemas.

Technical capabilities

Built-in Security & PII Protection

Requests pass through validation layers that block PII leakage, prompt injection, and malicious sources.

Drop-in SDKs

Official Python and JavaScript SDKs for integration with leading LLM providers.

Must watch videos

0:43

Tutorial0:43

Connect Tavily to Replit - here's how

In this quick how-to, we’ll show you how to connect Tavily to Replit and start using its search and extraction tools directly in your building workflow. With Tavily and Replit together, you can prototype web-aware agents, research tools, monitoring workflows, and other applications grounded in current information—without constantly switching tools. In this video: • Connect the Tavily MCP to Replit • Make Tavily’s tools available to Replit Agent • Test the connection with a real example Give Tavily a try for free: https://app.tavily.com/sign-up/?utm_medium=social&utm_source=youtube

See the difference in Muse results when using Tavily for research

3:13

Tutorial3:13

See the difference in Muse results when using Tavily for research

Our DevRel, Ricky, tested Muse on its own to search Costco.com for dairy- and soy-free snacks. After 25 minutes, results were still loading. Then, he connected Tavily to Muse, and tried again. In under 4 minutes, he had a list of snacks ready to peruse. Watch to see the results. Give Tavily a try for free: https://app.tavily.com/sign-up/?utm_medium=social&utm_source=youtube 0:00-0:016 Intro 0:17-1:26 Testing with Muse 1:27-2:09 Connecting Tavily to Muse 2:10-3:14 Testing Tavily in Muse

How Rox cut account research from a week to seconds with Tavily

1:56

Tutorial1:56

How Rox cut account research from a week to seconds with Tavily

See how and why Rox uses Tavily to stand up an infrastructure that now tracks millions of accounts and billions of contacts, cutting a sales rep's research time from weeks to running itself overnight. --- Rox agents are designed to help to put customers' pipeline on autopilot. But, to do so, the agents need to know what's happening right now at target accounts, and for users to trust the incoming info. That's why Rox needed a search and extraction tool its agents and customers could count on. Rox went through a few solutions before landing on Tavily as the answer. Costs were too large from one provider. Results were being flagged from another... "With Tavily, our agents can confidently rely on a web search provider that’s fast, cost effective and provides accurate data at scale." — Shriram Sridharan, Co-founder & CTO, Rox Watch the video, then read the full story: https://www.tavily.com/blog/rox-case-study/?utm_medium=social&utm_source=youtube Chapters: 0:00-0:27 Meet Rox 0:28-0:39 The challenge 0:40-1:20 Why Tavily 1:21-1:56 The outcome

Who uses it

Real use cases, no hype

RAG Grounding

Inject fresh, factual web context into LLM prompts to reduce hallucination.

Agent Web Access

Equip autonomous agents with real-time search, extraction, and site navigation.

Market Research

Automate gathering of competitor data, news, and financial filings at scale.

Content Aggregation

Build curated feeds or dashboards by crawling and extracting specific topics.

Fact Verification

Cross-reference claims against live web sources with structured data retrieval.

Competitive Monitoring

Track changes across competitor websites and industry sources automatically.

Academic Research Assist

Accelerate literature reviews by extracting and summarizing papers and articles.

What's great

Clean, structured content output ideal for LLM context injection.
Comprehensive endpoint suite: search, extract, crawl, map, research.
Direct integration with major LLM providers like OpenAI and Anthropic.
Generous free tier with 1,000 monthly credits to start.

Technical strengths

Blazing-fast 180ms p50 latency for web search.
Built-in safeguards against PII leakage and prompt injection.
Proven at massive scale: 300M+ monthly requests, 99.99% uptime.

Where it falls short

Pricing can scale unpredictably with high-volume research or crawl tasks.
Limited to web data; cannot access private or authenticated sources.
Credit system can be confusing for new users.
Documentation, while improving, can be sparse on advanced configuration nuances.

Technical limitations

Advanced features like `/research` may consume credits heavily.
No self-hosted option for full data control and air-gapped environments.

For technical folks

The deep-dive specs

FAQ

Includes technical Q&A

API credits are consumed based on the request type (search, extract, crawl, etc.). More complex or higher-volume requests use more credits. Usage details are in the documentation.