What We Shipped: May 2026
Snowflake Summit, new integrations (x402) and agent deployments across the ecosystem.
Meet us at Snowflake Summit
We’re kicking off June at Snowflake Summit 2026 in San Francisco, June 1-4. Join us and be a part of the conversation surrounding enterprise AI, agents, and connected intelligence! If you’ll be there, come say hi to the Tavily team at Booth 1613. Learn more about Snowflake Summit →
Now, here’s everything we shipped and everywhere we showed up in May.
Tavily is now on x402: Pay-Per-Call Web Search for Agents
Tavily is now live on x402, Coinbase’s HTTP 402-based micropayment protocol. This means AI agents can now access real-time web search and extraction without API keys, signups, or billing setup. Just runtime access, paid per request in USDC on Base.
Here’s how it works:
- Agent calls Tavily
- Receives 402 Payment Required
- Pays in USDC
- Gets instant results
One endpoint is already live at https://x402.tavily.com:
POST /search
- Real-time web search with ranked results, snippets, source URLS, and optional generated answers.
- $0.01 per call
Read the full launch details →
Tavily Research
More Control, More Precision
Our research agent already delivers powerful, accurate results - and now you've got more say in how it gets there. Two new controls:
- Domain inclusion & exclusion - point the agent at the sources you trust, or steer it away from the ones you don't
- Response length control - dial the depth up or down to fit the job
So your research lines up with your workflows, your requirements, and your sources of truth.
Explore the new parameter options in the research docs
Deep Research Now Accepts File Uploads
We’re excited to announce file uploads for Tavily Research: You can now attach (up to 20!) files to a research query, and the agent blends them right in with live web search. That means one run can pull from your internal data, past research context and of course - current public information from the web - all in one call.
We support PDF, Markdown, JSON, and TXT. We're especially excited about Markdown - the workhorse for agentic workflows:
- Agents write their research insights as an .md file
- Hand it to the next query
- Keep building instead of starting over
So your internal reports, contracts, briefs, and past research turn into context the agent can actually use.
Partners and Integrations
LangChain Interrupt Highlights
At LangChain Interrupt, Tavily was featured in a live demo showcasing a competitive intelligence agent built with LangChain Fleet and Nebius Token Factory. The system used Tavily for real-time web retrieval to continuously gather and structure cited insights across competitors. Throughout sessions and workshops, Tavily was a go-to for grounding agent workflows in fresh, reliable web context.
New Integrations to Keep You Moving Fast
ElevenLabs - Integrate Tavily with ElevenLabs through ElevenAgents to give your agents access to real-time web search. Learn more here.
Vellum - Use Tavily as the built-in web search provider in Vellum Assistant so conversations can be grounded in real-time web results. Learn more here.
Devin - Connect Tavily to Devin through the MCP Marketplace so Devin can search the web, read docs, and ground coding tasks with live web context. Learn more here.
Portkey - Plug Tavily into Portkey to inject live web search results into LLM requests, giving models up-to-date context before they respond. Learn more here.
Say Hello to our Newest Partner: ClickHouse
ClickHouse now integrates Tavily as a web search provider within LibreChat, enabling agents to pull real-time information from the web directly into chat and workflows.
With Tavily built into LibreChat, agents can search the internet and ground responses in up-to-date, relevant context.
Learn more about ClickHouse’s LibreChat →
Developer Perks Are Now Available on Three VC Accelerators
Tavily developer perks are now available through a16z Speedrun and Techstars, giving founders and builders in those programs easier access to Tavily as they build and scale AI products. The Collective support is coming soon.
On the Stage
We took the stage at IBM THINK to share how real-time web context powers agentic systems, and hosted a hands-on webinar with Nebius exploring how to build an expert agentic Slack bot using OpenClaw, Nebius Token Factory, and Tavily. See the full list of events and upcoming sessions here.
Community and Events
We Hit the Ground Running in May
We’ve been on the ground with builders across San Francisco, New York City, London, and Berlin this month. From hackathons to community meetups and developer sessions, our team enjoyed hanging out in all the places you were experimenting, shipping, and sharing ideas. Thank you to all who stopped by to say hello and chat with our team!
To see if we’ll be near you soon check out our events page →
Customer Stories
How Peerbound Found 5x More Accurate Data
Peerbound helps sales teams surface the right customer proof points at the right moment, but their underlying data layer was inaccurate nearly 80% of the time. By switching to Tavily, they replaced unreliable results with a real-time, grounded data source — moving from 2/10 accurate outputs to 10/10.
Peerbound’s 5x Data Accuracy Leap with Tavily’s AI Search API →
Loving Tavily? Tell the world
Real reviews from real users carry more weight than anything we could say ourselves. If Tavily has made your work easier, two minutes of your time could be exactly what someone else needs to find the right tool.
Connect with us
Questions or feedback on any of this? We're on Discord, X, and LinkedIn.