AI Assistants: Supporting Real-Time Knowledge Integration
JetBrains is known worldwide for its powerful development tools and IDEs, including IntelliJ IDEA, PyCharm, and ReSharper. The company’s mission has always been clear: help developers write better code, faster.
As part of its AI strategy, JetBrains set out to build JetBrains AI Assistant, bringing cutting-edge AI capabilities directly into IDEs. One recurring request from customers was ensuring the assistant could access current information from the web—especially developer documentation. That’s when JetBrains turned to Tavily.
Initially, the assistant struggled to answer questions about recent releases or framework updates. Developers quickly noticed the gap when asking straightforward queries like “What’s the latest version of Kotlin?” or “How do I do this in NumPy?” JetBrains explored several options, including Bing, Perplexity, and even building their own system. But each path came with high engineering costs, complexity, and maintenance—while lacking features critical for developer use, such as safe search, markdown formatting, and domain-specific filtering.
Tavily offered a smarter way forward. Its API integrated seamlessly with JetBrains’ products and supported two core use cases:
- General web search for up-to-date answers.
- Scoped documentation queries, where searches could be restricted to a single framework site like PyTorch or NumPy.
What could have taken months of engineering time—building safe search, clean formatting, and custom domain filters—came ready out-of-the-box. As one engineer noted:
Fast results, happier users
Complaints about outdated or missing responses disappeared, and JetBrains launched new features like scoped documentation queries with minimal development effort.
The adoption also reflected JetBrains’ bottom-up culture: developers, who were already familiar with Tavily, advocated for it. Leadership approved based on that trust and its proven technical fit.
A smooth collaboration with long-term potential
Beyond the technical benefits, JetBrains emphasized the smooth collaboration. Tavily’s team responded quickly to requests, including adding crawling support when JetBrains needed it most. As Alexandra Charikova put it: “We enjoyed how smoothly the Tavily team worked with us—from zero to launch.”
Looking ahead, JetBrains is considering other opportunities where Tavily will play a key role in the platform.
As Alexandra Charikova from JetBrains said: