Semantic search across user's context

Performs semantic search across the user's connected integrations and returns raw matches from messages, meetings, and calendar events.

Billing: This endpoint is metered. Each call consumes AI query credits.

Use this when you need:

• Raw communication data without AI summarization
• Full conversation threads plus matching meetings/events
• To build your own UI or processing on top of search results

Response structure:

• key_messages: Messages most relevant to the query
• conversations: Conversation threads containing the key messages
• key_meetings: Matched meetings (empty array when no meetings match)
• key_events: Matched calendar events (empty array when no events match)
• Every returned record includes external_id (partner-provided ID from ingest, or null)

Parameters:

• query: The search query (semantic, not keyword-based)
• integration_ids: Optional filter to specific integrations (opaque IDs like int_xxx)
• conversation_ids: Optional filter to specific conversations (opaque IDs, e.g., "conv_xxx")
• after_date: Optional date filter (YYYY-MM-DD)
• before_date: Optional date filter (YYYY-MM-DD)
• sender_domains: Optional filter by sender email domains (e.g., ["acme.com"])
• entity_ids: Optional filter by entity IDs (UUIDs)
• retrieval_mode: Optional retrieval strategy — "fast" (lowest latency, default), "auto" (adaptive), or "deep" (full pipeline)
• expand: Optional comma-separated fields to include (e.g., "body_text,summary,description")

Expand parameter: By default, expandable detail fields are omitted to reduce payload size. Use expand=body_text for message content, expand=summary,short_summary,action_items,attendees,meeting_link for meetings, and expand=description,location,html_link,event_link,attendees for calendar events.