ToolFlock

Exa

Exa is an AI search engine and API that returns real-time web data for AI agents, chatbots and research workflows. It pairs a large public web index with page-content fetching and deep research endpoints.

#27 in AI AssistantsAI Search EnginesFree plan

Who it's for

Best for

  • Developers adding web search to coding or chat agents
  • Analysts doing unstructured account and topic research

Not for

  • Sales teams building verified company and contact lists at scale
  • Searchers who need a consumer search interface rather than an API

Features

  • Search API across web verticals

    A search endpoint returns ranked results from a large public web index, with verticals covering full web, news, company, research and financial sources.

  • Contents API with highlights

    The Contents API pulls page text and uses highlights to pass concise excerpts to models instead of full pages.

  • Agent API with latency modes

    An Agent API runs deeper research, with Lite, Auto and Deep modes trading speed, from roughly 450 milliseconds up to about ten seconds.

  • Monitors and crawler tools

    Website crawler tools fetch page content, and a separate Monitors API watches pages for changes alongside the core search endpoint.

  • MCP, Claude and ChatGPT connectors

    MCP server access, a Claude Connector and a ChatGPT plugin connect the search to existing assistant and agent tools.

What users say

Pros

  • Semantic retrieval returns the material asked for without the constant prompt re-optimisation that literal keyword search demands.
  • Article content fetching is cheap, and a run of queries draws only a small slice of the prepaid balance.
  • Unstructured account research — articles, opinion, context — returns solid material and stays in use for that layer.
  • As the web search layer behind coding and chat agents, it returns usable context for the agent to reason over.

Cons

  • Structured company lists come back stale and unfiltered, with headcount pulled from old articles and no deduplication between entries.
  • Repeated trial runs produced unusable output that read like generic AI slop.
  • Content fetching repeatedly fails with crawl errors and live-crawl timeouts instead of returning the page.
  • Entity lookups returned six results with only two correct and missed known blogs entirely, and a unique real name surfaced no linked pages.

A selection of the review sources we used:

Summarized from public user reviews and community discussions. View details

Global Traffic Rank

Top 500k

Alternatives

ExaBrave SearchPerplexityKagi
Rank#27 in category#2 in category#7 in category#14 in category
Free planYesYesYesYes
ProSemantic results, cheap contents, strong research, agent searchindependent index, privacy, sourced AI answers, bangs, free APIClear, current answers backed by sourcesClean, ad-free results, customizable site ranking
Better forEveryday consumer web search using shortcut bangs and custom GogglesInteractive conversational search with synthesized answers and cited web sourcesAd-free consumer search with custom domain reranking and website blocking
Weaker atProcessing long agent queries and returning full page contents directly via APIProgrammatic page extraction, token highlights, and web monitors for agent pipelinesExtracting clean page text and highlights for AI agent pipelines