Exa adds a lite search type on October 19, 2026: $4 per 1k requests, and fast, instant and deep-reasoning become legacy
Exa /search gets type lite at $4 per 1k requests (up to 10 results), cache only. fast, instant, deep-reasoning are now legacy.
Your keys
0 of 1 ready
Create each key on its own page, paste it below, then copy the finished lines into your project's .env file. What you paste stays in this tab. It is never sent to Wirestack or saved.
- Exa API keyEXA_API_KEYGet key
Create a key. New accounts get free credits
What you get
On October 19, 2026 Exa added a new lite search type. You set type: "lite" on /search. It costs $4 per 1k requests with up to 10 results, plus $1 per 1k results above 10. It serves page content from Exa's cache and never fetches pages live. Parameters it does not support return a 400 error that names them.
Search responses now include a warnings array for request fields that have no effect, and a notices array that explains how results were produced. The fast, instant and deep-reasoning types are now legacy. They still work at their current prices. Exa suggests lite for instant, lite or auto for fast, and deep or Exa Agent for deep-reasoning.
Prerequisites
Steps
- Find every call that sets
typetofast,instantordeep-reasoning. - Switch the ones that can use cached content to
lite, following the lite guide linked from the changelog. - Remove any parameter that
literejects with a400.
Test it
Run one query with type: "lite" and check that results come back and the warnings array is empty.
Not confirmed
- A shutdown date for the legacy search types.
- How fresh the cached content used by
liteis.
