Introducing Web Search Agents: expert level web search that self-learns your domain to deliver perfect web context, while burning less tokens. Generic web search tools start from scratch every time they run. Web Search Agents compound their expertise to adapt web search and crawling to your use case —... finding the exact data you need, more efficiently. Learn more below in our blog:show more

Unlike generic web search tools that start from scratch when they search the web, Web Search Agents remember what worked last time and templatizes recurring searches (i.e. find my brand mentions on social every day). This means less search requests, executed more efficiently…requiring less tool calls and token spend for you. Learn more about Nimble's self-learning here:

Nimble Javascript rendering enables your agent to access any data on the web, while real-time browsing goes deep into pages that might not turn up in search results.

Here's how to get started with Web Search Agents in five different verticals: Legal - research regulations and case law: Social - identify targeted influencers: GTM - run company research: Coding - research coding docs: Finance - evaluate analyst predictions against actuals: Don’t see your use case? Tag us and we’ll build one for you.

How to integrate Web Search Agents with your stack: Langchain: Open AI: Anthropic: Vercel: LlamaIndex: Hermes: See more here:

Web search vendors love public benchmarks like BrowseComp, DeepResearch Bench, and DRACO. We can run those benchmarks too, but its not the best way to evaluate web search for your agent (see our next comment),

Our question for the industry...why evaluate web search using benchmarks that don't reflect the kinds of queries your agent will make? We think the smarter approach is to evaluate web search against vertical benchmarks that reflect your domain. Notice how Nimble's accuracy improves every run as it learns the use case.

Love this. Compounding expertise less tokens, better results

Pleasure working together on the launch video! 🔥

Well done, been building with these for a couple of days. my favourite detail: mark a field nullable, and an unpublished figure comes back as null with a citation proving it isn't published

Thanks @sebuzdugan !!

This is huge Web Search Agents that actually get smarter with your domain Less token waste better results The future of search right here

Go Nimble!

Expert level search. Go deep!

Congratulations.. This is powerful 💪

Thanks @ai_for_success !!! Can't wait to see what you build with it 👀

I built my own custom agent builder which can build any agent with schema + you can call it and use whenever you want.. Posted and made the repo opensource so others can try... I had fun building it.. It's really good.

Smarter search can save a lot of time and effort.

Self-learning web search that compounds expertise and burns less tokens is exactly what we needed!

Great innovation making AI search more specialized and efficient

Amazing Smarter search that learns, saves tokens, finds everything.

That's right!

The compounding approach is promising. Love this.

Nimble launches expert web search agents