Insight O' Mate in one page

Insight O' Mate turns plain-English questions into queries for your NoSQL databases, then runs those queries on your database and shows you the results — without streaming your data to anyone else's cloud.

It ships as a desktop client that talks to two things:

• A NoSQL database of your choice — MongoDB, Redis, Firestore, or DynamoDB
• A stateless NLP engine we host (or that you self-host), which only ever sees your prompt and your field names — never your rows.

Why local-first

Most "chat with your database" tools upload your data to a third-party LLM. We think that's a non-starter for any team whose database contains anything users would care about.

The architecture is deliberately boring:

1. You type a question in the desktop app.
2. The app fetches the schema of your collections (field names + types — never values) and sends it along with your prompt to the NLP engine.
3. The engine returns a structured intent object (an AST), not the final database query.
4. The local desktop sidecar takes that intent and translates it into the actual database syntax (a MongoDB pipeline, Firestore query, etc.) on your machine.
5. The app runs that query directly against your local DB. Results are rendered in the app, and nothing persists to our servers.

See Privacy model for the full what-leaves-and-what-doesn't.

What's in the rest of these docs

• Getting started — install, connect, run your first prompt.
• Connecting databases — supported deployments, read-only roles, connection strings.
• Privacy model — exactly what leaves your machine.
• API reference — the /analyze endpoint if you want to script it.

Who this is for

• Solo founders who want answers from their own database without standing up a BI stack.
• Operators at small teams who don't want to wait on data engineering.
• Security-conscious devs who've been burned by "just paste your schema into our chat box" workflows.

It is not a replacement for a proper warehouse. If you're running terabyte-scale analytics, use ClickHouse or BigQuery — we're aimed at the operational database tier.