Search

One hybrid (semantic + keyword) search across the workspace's content — conversations, project files, and memory — scoped to what the caller may access.

Search

GET/v1/workspaces/{workspace_id}/search —

query params: q (the query), sources (which content types to include — conversations / files / memory), and limit.

bash

curl -G https://nexevo.ai/v1/workspaces/$NEXEVO_WORKSPACE/search \
  -H "Authorization: Bearer $NEXEVO_API_KEY" \
  --data-urlencode "q=onboarding checklist" \
  --data-urlencode "limit=10"

Results are filtered to the caller's access (their conversations, accessible projects, the workspace memory) — search never crosses tenant or permission boundaries. For pure long-term recall, query Memory directly.