Web search for AI

AI web search has become a confusing mess. OpenAI, Anthropic and Google have search built into their models, while Exa, Tavily, Firecrawl, Jina and Liner all offer overlapping versions of search, extraction, crawling and research. In this video, I break down what each provider is actually good at, what its Convex component adds, and when I think you should skip them all and use the model's built-in search instead. My basic rule: - If only the model needs a current answer, use built-in web search. - If your application needs to inspect, store or reuse the results, look at Exa, Tavily, Jina or Liner. - If you need to acquire pages or crawl a site over time, use Firecrawl. - If your application already owns the documents, use Convex full-text or vector search. This isn't a search-quality benchmark. It's a practical look at which tool fits which job. The five providers: - Exa for semantic discovery and structured research - Tavily for general-purpose agent search - Firecrawl for scraping and durable site crawling - Jina for turning URLs into clean, reusable content - Liner for academic search and cited research The components are open source, but you still pay the provider's API costs and normal Convex usage. Disclosure: Exa supplied Convex with API credits and built a demo after we started discussing this video. They had no control over my conclusions. Convex components Full directory https://www.convex.dev/components Exa https://www.convex.dev/components/exalabs/convex-exa Tavily https://www.convex.dev/components/tavily/convex-tavily Firecrawl https://www.convex.dev/components/firecrawl/firecrawl-convex Firecrawl Scrape community component https://www.convex.dev/components/firecrawl-scrape Jina https://www.convex.dev/components/convex-jina Liner https://www.convex.dev/components/liner-convex Demo source Exa https://github.com/exa-labs/exa-convex Tavily https://github.com/tavily-ai/convex-tavily Firecrawl https://github.com/firecrawl/firecrawl-convex Jina https://github.com/adhishthite/convex-jina Liner https://github.com/mjkang-estrella/liner-convex Provider pricing Exa https://exa.ai/pricing?tab=api Tavily https://docs.tavily.com/documentation/api-credits Firecrawl https://www.firecrawl.dev/pricing Jina https://jina.ai/reader/ Liner https://liner.com/developers/docs/pricing Native Convex search Full-text search https://docs.convex.dev/search/text-search Vector search https://docs.convex.dev/search/vector-search Built-in model search OpenAI https://developers.openai.com/api/docs/guides/tools-web-search Anthropic https://platform.claude.com/docs/en/agents-and-tools/tool-use/web-search-tool Google https://ai.google.dev/gemini-api/docs/google-search OpenRouter https://openrouter.ai/docs/guides/features/server-tools/web-search #Convex #AIAgents #WebSearch