product

missionsagent-driven promotion campaigns with adjustable autonomycontent & draftsAI-generated content reviewed before it shipsinbox & engagementsunified inbox for mentions, drafts, and AI decisionsCRM & contactsrelationship management with AI-powered outreachresearch & recordsstructured data gathering on autopilotworkflowsvisual automation graphs — no code requiredprojects & docswork planning, docs, and standups — all native

solutions

for startupsyour marketing team is an agentfor agenciesscale clients, not headcountfor creatorsgrow your audience while you createfor enterpriseautonomous marketing with full controlview all solutions
platformmodelspricing>docs

Search

Global search finds anything in your workspace from a single box — subjects, contacts, tickets, drafts, knowledge, skills, missions, projects, documents, conversations, engagements, decisions, and team members — ranked together in one list. It's available three ways: the Cmd+K palette in the UI, the search.global API endpoint, and the search.global agent tool.

How it ranks

Each query runs three signals in parallel and fuses them (Reciprocal Rank Fusion):

  • Full-text — Postgres tsvector matching on titles and bodies.
  • Fuzzy — trigram similarity (pg_trgm), so typos and partial words still match.
  • Semantic — vector similarity over embeddings, so a paraphrase finds the right thing even with no shared words.

You can scope a search with mode:

  • keyword — full-text + fuzzy only (fastest; no embedding call).
  • semantic — vector only.
  • hybrid — all three, fused (the default).

Permissions

Results are always workspace-scoped and RBAC-gated: you only ever see hits for things you're allowed to see. Credentials, secrets, API keys, and agent-internal memories are never indexed.

From the UI

Press ⌘K (Ctrl+K on Windows/Linux) anywhere to open the search palette. Type to search; results are grouped by type. Use the arrow keys and Enter to jump straight to an entity. (Docs search lives on ⌘/.)

From the API

POST /api/invoke/search.global
Content-Type: application/json
 
{ "q": "kickstarter backers", "mode": "hybrid", "limit": 20 }

Returns a ranked list of lightweight hits, each with a deep link:

{
  "items": [
    { "type": "contact", "id": "con-…", "title": "Ada Lovelace", "snippet": "…", "score": 0.03, "url": "/contacts/con-…" }
  ],
  "total": 12
}

Optionally pass types (e.g. ["contact","subject"]) to restrict to specific entity types, and offset to page.

See the search.global reference for the full input/output schema.

How the index stays fresh

When you create or edit an entity, its keyword and fuzzy match update immediately. The semantic (vector) embedding is filled in shortly after by a background reconcile pass, so brand-new items are findable by keyword right away and by meaning within a minute or two.