Are you heading to @LangChain Interrupt26, NYC? Stop by our booth, say hello, and join the closing reception we're sponsoring from 5:00-6:30pm. During the event, you'll hear keynotes from industry leaders, have access to hands-on workshops, and see LangChain agents using real-time web retrieval with Tavily. After the event, we're hosting an exclusive dinner with Nebius, where our Founder and CEO, Rotem Weiss, and @nebiusai's Developer Relations Advisor, Rob Spectre, will get the conversations started on what it takes to move agents from demo to production. Request to join the dinner:

You can now bring real-time web search directly into your @Replit workflow via Tavily 😎 Our DevRel, @tavily_rick, put together a quick demo showing how to connect Tavily to Replit and start using it. Watch below. And if you’re in San Francisco on October 9, Replit will be co-hosting Builders & Brews with us. We’ll be bringing together developers, founders, and AI builders for an evening of agents, coffee, and hands-on building. ☕ Join us at Builders & Brews:

We're excited to introduce and be a part of Agentic AI, a free hands-on course from @Nebius_Academy and @NVIDIA where you get to build your own deep research agent that searches the web, reasons, delegates to subagents, and gets deployed. The whole course takes about four hours, and curriculum includes: Module 1: Intro to managed inference Module 2: Agentic web search (preview video by @lakshyaag attached) Module 3: Orchestrating deep research Module 4. Evaluating Deep Research Module 5. Making persistent agents Module 6. Agent deployment Enroll and run through the modules for free, get a digital badge, and then take the Agentic AI Builder Certification exam for $1. Get started:

We sat down with Shriram Sridharan, Co-founder & CTO at @rox_ai, to hear what it took to add the right web search provider to the GTM agents their customers depend on. The Rox team found with their original provider, customers weren’t getting up-to-date results, and costs were moving up. That’s when Rox turned to Tavily. Hear how Rox stood up an infrastructure in two weeks that now tracks millions of accounts and billions of contacts, cutting a sales rep's research time from weeks to running itself overnight:

We're co-hosting Alpine Agentic Search, an AI community meetup with @nebiusai in Zurich, Switzerland on Wed. October 7. 𝗜𝗳 𝘆𝗼𝘂'𝗿𝗲 𝗶𝗻 𝘁𝗵𝗲 𝗮𝗿𝗲𝗮, 𝗰𝗼𝗺𝗲 𝗷𝗼𝗶𝗻 𝘂𝘀! We'll be talking about everything from low-level search primitives on modern GPUs, through advanced retrieval and ranking, to the latest developments in agents for search workloads. You'll hear four technical talks, including one from @NVIDIA, and have time for networking, drinks, and bites. Request to join:

The Nebius AI Builder Program is a developer community where you can try the newest models, start a build with working code, and compare notes with Nebius, partner, and fellow engineers. 𝗧𝗮𝘃𝗶𝗹𝘆 𝗶𝘀 𝗽𝗿𝗼𝘂𝗱 𝘁𝗼 𝗯𝗲 𝗮 𝗽𝗮𝗿𝘁 𝗼𝗳 𝗶𝘁. Come join the free program, and get: - Credits and discounts from Nebius Token Factory, LangChain, Toloka, Tavily, and more. - Access to an engineering community including office hours and Discord - Cookbooks and working examples - Free courses and discounted certs See what it's all about.

Check out this simple and quick walkthrough on using agent-owned secrets in LangSmith via Tavily. Also shown are a user-owned MCP OAuth connection with Linear and user-owned OAuth in a custom app with GitHub. 👏

The Private Memory Layer for AI now reads the web with @tavilyai. Anuma agents pull live pages through Tavily Extract, with failover across providers so a slow page never stalls an answer. Available now in @AnumaAI. 290,000+ wallets.

We’re excited to welcome @tavilyai as a sponsor of The AI Conference 2026. Tavily is the web access layer for AI agents, providing real-time search, content extraction, crawling, and research through a secure API. Built for production at scale, Tavily helps developers connect AI agents to up-to-date web context while adding safeguards for security, privacy, and content validation.

The Managing Director, ERPM Data and AI Technology at @BMO, Ashutosh Sinha, realized that the manual research their analysts were running for AML either required adding more people or automation to keep up with demand. So, Ashutosh turned to Tavily for automation.  BMO technology and the Tavily team worked together to build a retrieval layer for secure web search and faster risk determination that allows analysts to move through more reviews, with confidence in the findings.  This was the first production use case with Tavily for BMO, and it won't be the last. Read the full story, and see why Tavily fit right in with BMO's security review:

Your @NanoClaw_AI agent just got a new, live view of the web. Tavily Search and Extract are now available in NanoClaw, and there's no API key required to start. With the Tavily integration, your persistent, isolated NanoClaw agents in Slack, Telegram, Discord, WhatsApp, and the CLI, come with: - Search and Extract for current, source-backed answers - A keyless start, then connect a Tavily key when you need more capacity (the agent never sees the key) - Reliable returns you can use for developer research, scheduled intelligence briefs, or a chat assistant that reads a URL and answers with receipts Demo below. You can run the /add-tavily-tool in NanoClaw providers that support MCP, and install it on multiple agent groups. Here’s how:

When you're fighting cybersecurity threats running on automations that cause 80K-100K alerts per day and your analysts need to review quickly, you have to act fast. So, #BellCyber#'s Chief Technology and AI Officer, Jawed Ahmad, chose to add automations to fight automations. After testing mutliple providers, Ahmad chose Tavily to provide the research layer their analysts could count on for fast, source-backed findings. The agent decreased investigation time from 30 minutes to 5 minutes, removing the research bottleneck and freeing up analysts to get through far more alerts in a day. Read the full story: @stratejm

What does building your own AI agent search layer truly costs when compared to using a purpose-built solution? Our DevRel, Ricky Johnson, did a side-by-side comparison, wrote out his findings, and explains how you can do the same to determine which is best for your needs. Ricky covers: - Returns from self-hosted search vs a purpose-built product - Costs in time vs dollars - When to choose one over the other Give it a read, then try a test of your own:

Yesterday, we teamed up with @jpmorgan and @Nebiusai for an AI-Native Founders Lunch and found out that even though everyone is building very different things, they're all hitting the same kinds of challenges. We dug into the hard questions, like, what they're refusing to build, even though customers keep begging for it, and got some great, honest answers. Thank you to everyone who joined us at J.P. Morgan's NYC office for a delicious lunch and thoughtful conversation. And thank you to J.P. Morgan and Nebius for building this with us. Same energy next time! J.P. Morgan #InnovationEconomy#

What happens when you power investment research with Tavily? Five workstreams kick off in parallel, pulling info and sources from across the web (see the video). When done, the findings come together in a meeting-ready brief, opening with an executive summary, then grouping results across policy, consensus, and disagreements. Financial agents can use Tavily to safely: ➤ Surface filings, news, and earnings calls in real time ➤ Research thousands of companies at scale ➤ Generate grounded reports with source citations See how your team can use Tavily for investment research; KYC, AML, and compliance; and risk management:

Helping to put your pipeline on autopilot is what @Rox_ai agents do. But that requires them to know what's happening right now at your accounts, and for you to be able to trust the info coming in. That's why Rox chose Tavily to run search and extraction. 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_s, Co-founder & CTO, Rox See how Rox stood up an infrastructure with Tavily in two weeks that now tracks millions of accounts and billions of contacts that cut a rep's research time from weeks to running itself overnight:

Not all search APIs solve the same problem. And in regulated industries like finance, you need to be sure the systems you're building can be trusted to meet compliance, reduce risk, and pass audits. In our blog, we break down three questions you should ask before choosing a web search API for finance, and highlight three examples of financial institutions that are putting Tavily to work for different uses: 1. As a retrieval layer under an investment research agent 2. For preserving source-level attribution in AML workflows 3. For source-attributed context their forward-looking risk models can use directly Here's a quick guide for choosing the right retrieval layer for agents in investment research, AML, and risk workflows:

Will you be in NYC on September 24? Come connect, learn, and explore the future of agents with us at Interrupt, the Agent Conference by @LangChain. We're incredibly excited to sponsor Interrupt, and would love to see you there in person. Here's more info:

Because search quality is often the limiting factor for AI agents, we spent the past several months rebuilding core parts of Tavily /search, including evidence ranking, contradiction handling, and index coverage.  Using the same retrieval budget, reader model, and grading pipeline across providers, 𝗧𝗮𝘃𝗶𝗹𝘆 𝗶𝘀 𝗻𝗼𝘄 𝗡𝗼. 𝟭 𝗶𝗻 𝗮𝗰𝗰𝘂𝗿𝗮𝗰𝘆 𝗼𝗻 𝗦𝗲𝗮𝗹𝗤𝗔-𝗵𝗮𝗿𝗱, 𝗦𝗲𝗮𝗹𝗤𝗔-𝟬, 𝗮𝗻𝗱 𝗦𝗶𝗺𝗽𝗹𝗲𝗤𝗔 𝗩𝗲𝗿𝗶𝗳𝗶𝗲𝗱.  See how and why these improvements help your agents retrieve better evidence and produce more accurate answers: