Sentra

About Sentra

Sentra is a unified memory layer — a company brain shared by a team and its AI agents. It captures interactions, decisions and drift and writes them into one queryable graph exposed over REST or MCP, so what you teach one agent every agent remembers. The technical argument is specific: vector search returns what is close, not what is correct, because embeddings are stored at write time and structure is guessed at query time, which means every request re-crawls Slack, email and docs to rediscover what a term means inside your organisation. Sentra instead resolves semantics at ingestion and builds the graph on demand at query time against an ontology unique to your company, treating meaning as a primitive rather than a side effect. It also argues that most tools capture artefacts — the CRM record, the Jira ticket, the Confluence page — and therefore capture only the output of a decision, not the interaction that produced it. In practice the product covers meeting recording and notes, pre-meeting briefs, deep research across all connected data, a commitment tracker, a risk radar and connection intelligence, plugging into Slack, Gmail, Calendar, Outlook, GitHub and Linear, and into Claude, ChatGPT, Cursor, Perplexity, Codex and Windsurf.

Key Features

Sentra Pros & Cons

Tags