Exa

Use Exa for web search, page content, similarity discovery, answers, asynchronous research, and structured Websets. The integration provides 23 tools: 22 retained tool keys plus native team context. API keys are sent only to https://api.exa.ai through x-api-key; redirects are refused. Team context identifies a team, not a person.

Search and content operations can consume credits. Research jobs and Webset searches or enrichments retain state; monitors create recurring billable activity. A failed receipt can follow a successful effect. Inspect the exact native resource before retrying; cancellation and deletion do not promise refunded credits or erased history.

Websets use native IDs and cursor pages. Item URLs come from native properties, enrichment results retain their native array/null values, and the historical enrichment map is keyed by native enrichment ID. Title and external ID can be absent. Current updates support title and metadata; changing external ID is refused. Monitor creation requires a search count and a supported five-field schedule; the historical refresh input is retained with actionable refusal.

Research remains supported by the current official JavaScript SDK at /research/v1. Historical Webset export routes are retained as unverified compatibility: their presence in this integration is not evidence of current provider availability. They return validated legacy metadata only; file ownership, expiry, delivery and renewal are unverified. Use Exa directly for files until a current native contract is established.

The integration has no triggers. Live provider acceptance has not been verified. Local verification uses controlled responses and denies external transports; the active private suite has read-only native observations and gates paid or retained effects before invocation.

License

This integration is licensed under FSL-1.1.