Research reports
The new async Research API plans, gathers, and computes a full report from a single question, then returns a reproducible result you can re-run. It is available now on every paid plan.
// CHANGELOG
Changelog
New endpoints, faster responses, and product updates. Follow along for what is shipping on the Scout API.
June 2026
A citation behind every claim
Research answers now carry per-claim citations, so every fact links back to the source it came from and the whole report can be audited line by line.
New plan lineup
Plans now span Builder, Scale, Pro, Business, and Premium, with Enterprise for custom volume. Each step up raises your monthly credits, lifts your rate limits, and lowers your overage rate.
May 2026
Faster web search
Search responses are noticeably quicker across common queries, and multi-page result fetches are more reliable when you ask for more than the first page.
Cleaner subpage discovery
Extract now follows real content links instead of navigation, login, and footer noise, so structured results from multi-page sites come back tighter and more on-topic.
Partial failures surface clearly
When one part of a fan-out request fails, the response now tells you exactly what succeeded and what did not, instead of failing the whole call silently.
April 2026
Bring your own output schema
FindAll honors a JSON schema you supply, so generated lists come back as typed rows in the exact shape your code expects, on both the sync and async paths.
Enrich merges instead of replacing
Enriching an existing list now merges new fields into each row and processes the full set, so earlier data is preserved and nothing gets truncated.
Streaming tool calls
The web-grounded Chat endpoint streams tool-call deltas as they happen, so you can render progress in real time instead of waiting for the full response.
March 2026
Brand and logo endpoints
Look up a company by name, email, or ticker and fetch its domain, brand details, and a high-quality logo, with a clean color-matched fallback when no logo exists.
Scheduled queries with webhooks
Monitors run a search on a schedule and post to your webhook the moment results change, turning the live web into an event stream your product can react to.
Credits refunded on failure
Every endpoint now zeroes or refunds credits when a call does not deliver a result, so you are only charged for work that actually lands.