Conrad Philmore

Closed sandboxes aren’t the future. Portable skills + real-time retrieval are. Tavily enables both without locking you into one model or API. Tavily.com

We’re now live on @mastra with a native Tavily integration with doc updates, as well as on mastracode with new Tavily native imports. Now instead of manually stitching APIs together, your agent can call Tavily search, filter results, and extract structured data just like any other built-in tool. mastra.ai/reference/tool…

Tavily Search gives you the right results. No hallucinations. tavily.com

⛷️ Building AI agents in @Snowflake? New developer guide: a full due diligence + investment research agent using Tavily's Search API for real-time web grounding. Check it out here: snowflake.com/en/developers/…

Hilarious!

We built Tavily for moments like this. Now it’s inside OpenClaw and agents can search, ground, and respond with real context. Your turn to build something wild. github.com/openclaw/openc…

Tavily is officially sponsoring our first-ever tradeshow and we’re heading to #HumanX in San Francisco 🌆 If you’ll be there, come meet the team at Booth #308 so we can connect. We’re bringing a few things worth stopping by for: 🎧 Apple AirPods Max headphones giveaway 👕 Tavily merch (no landfill tees, promise) 🎓 A live masterclass with @nebiusai ⚡ Live demos of Tavily powering real-time AI search Swing by, say hi, and let’s talk about what you’re building.

With Tavily Search + @nvidia AI-Q + @LangChain, you can build deep research agents that combine private data + real-time web intelligence, all in one workflow. developer.nvidia.com/blog/how-to-bu…

@AdolfoUsier @tavilyai Couldn't agree more. I was really excited to see this news.

@tavilyai Tavily grounding cuts hallucination way down in production - solid add to the stack

Curious about methods to reduce caching cost? Our friends over at @awscloud have some ideas. Check it out! aws.amazon.com/blogs/storage/…

Always interesting to see the cool stuff AWS is working on.

Building coding agents isn’t the future, it’s already happening. Join us for a live, practical walkthrough with a Tavily Forward Deployed Engineer @EvanRimer on how teams are shipping production-grade coding agents today. Then see how @jetbrains integrated Tavily to power their new agent, Junie. Bring your messy ideas, weird edge cases, and big ambitions. Leave with something you can actually build. Let’s get into it, register here: tavily.com/webinars/build…

Have you taken the Tavily Certification yet? It might be time to take your Agent’s search skills to the next level!

Most search APIs return links. Tavily returns usable context: ranked sources extracted content clean inputs for LLM reasoning Built for agents, not browsers.

AI agents are already operating inside the terminal. From OpenClaw to Claude Code to Codex, agents can plan, execute, and iterate through commands. The next step is making sure they can reliably access and use real-time web data. That’s where the Tavily CLI comes in. Full breakdown ↓ tavily.com/blog/tavily-cl…

Tavily is now a web search option in @openclaw . Your agents get: real-time web results high-signal context reliable grounding Built for agents, not browsers. github.com/openclaw/openc…

@tavilyai Been waiting for this, great update!

🧪 We’ve been working on a major improvement to our image retrieval and the results will help you find better images faster. With this update Tavily users will see: • Improvement in retrieval accuracy • Simpler multimodal pipelines • Increase in relevant image matches What does that mean in practice? Developers can now build multimodal apps and agents without the extra complexity. Text and images come back together in a single response, so there's no need for a separate image search or matching layer on top. Stay tuned! tavily.com/blog/better-im…

Tavily helps teams move from question → credible sources → clear answers in one streamlined search workflow. From due diligence to feature planning, good decisions start with good inputs. Tavily.com

AI agents are only as good as the context they see. Reranking prioritizes high-quality sources before results ever reach the model, reducing hallucinations and giving teams answers they can actually trust.