Insight O' Mate

About Insight O' Mate

Insight O' Mate is a privacy-first AI database copilot and Text-to-NoSQL platform. It lets developers, analysts, and business users query their databases and spreadsheets using plain English — without writing complex query syntax — while keeping all data and credentials on their own machine.

What It Does

Insight O' Mate translates natural-language questions into database queries and executes them locally against your data. You describe what you want to know, and the product generates the corresponding query for your database, runs it directly from your desktop, and shows you the results.

No data upload. No cloud processing of your records. No pasting proprietary schemas into a public AI chat interface.

Supported Databases and Data Sources

Insight O' Mate currently supports:

  • MongoDB — aggregation pipelines, find, countDocuments, and distinct via Atlas, replica sets, and standalone instances
  • Firebase Firestore — collection queries with where filters, orderBy, limit, and count aggregations via the Admin SDK
  • Redis — Hashes, Sorted Sets, Lists, Sets, and Streams via HGETALL, ZRANGEBYSCORE, LRANGE, SMEMBERS, XREVRANGE, and non-blocking SCAN
  • Amazon DynamoDB — Query and Scan operations with automatic KeyConditionExpression, FilterExpression, ExpressionAttributeNames, and ExpressionAttributeValues
  • Excel (.xlsx, .xlsm) and CSV — in-memory aggregation, filtering, grouping, and cross-sheet joins without formulas or data upload

Planned databases on the roadmap: Cassandra, Elasticsearch, Neo4j, and Couchbase.

Who Built It

Insight O' Mate is a product of Lunamic. We're a small team building developer tools. The product started because the safest way to query a database with an AI assistant — pasting your schema into a public chat interface — is also a privacy and security risk for anyone handling sensitive data. We set out to build the alternative: a local-first architecture where the AI never sees your data.

Why It Exists

Most teams that work with NoSQL databases face a recurring friction: the query syntax is complex, documentation is scattered, and hand-writing aggregation pipelines or expression attribute names takes time even for experienced engineers. Data analysts without deep database knowledge depend on engineers for every data question. Business teams wait days for simple answers.

The standard solution — copy-pasting schemas or data samples into ChatGPT or Claude — works, but it sends confidential structure and sometimes real records to a third-party cloud. For teams working with healthcare data, financial records, or proprietary customer data, that tradeoff is unacceptable.

Insight O' Mate solves this by keeping query generation as the only cloud interaction. The prompt and field names travel to a stateless NLP engine. Everything else stays on your machine.

How It Works

  1. Connect your database — paste a connection string or upload a service account file. Credentials are stored in your OS keychain, never transmitted.
  2. Ask a question in plain English — for example: "Find all customers who placed more than 3 orders this month."
  3. The NLP engine receives only your question and the collection or table field names — never your document values or credentials.
  4. A query is generated — a MongoDB aggregation pipeline, a Firestore SDK call, a Redis command, a DynamoDB expression, or an in-memory spreadsheet operation.
  5. The query runs locally — the desktop client executes it directly against your database from your machine.
  6. You see the results — inspect the generated query, copy it into your codebase, or ask a follow-up question.

Privacy Philosophy

Our privacy architecture is not a feature we added — it's the reason the product exists. The precise boundary is:

What leaves your machine: your plain-English prompt and collection or field names only.

What never leaves your machine: document values, row data, query results, connection strings, credentials, and API key contents.

The NLP engine is stateless. It never logs prompts or results, never trains on your data, and never stores session state. You can verify this yourself: enable IOM_LOG_ENGINE_REQUESTS=1 and inspect every outbound request as plain JSON. Read the full security and privacy model and the technical privacy model docs.

Who It Is For

  • Software developers and backend engineers who build on MongoDB, Firestore, Redis, or DynamoDB and want to explore or debug data without hand-writing every query
  • Data analysts who need to ask questions about database content in plain English without learning MQL, Firestore SDK syntax, or DynamoDB expression language
  • Business analysts who work with Excel and CSV data and want to replace complex formulas with natural-language questions
  • Privacy-conscious organizations that cannot send database structure or records to cloud AI services

How We Make Money

Paid plans — see pricing. The free plan is genuinely free because a generous free tier is the best way to let people try the desktop app. We don't sell your data. We can't, because we don't have it.

  • Free: 20 database analyses per day
  • Pro ($29/month): Unlimited queries, priority queue, advanced export
  • Ultra ($79/month): Dynamic dashboards, dedicated NLP instance, VIP support

What's Next

We are in public beta. See the full product roadmap. In rough order:

  1. Cassandra and Elasticsearch adapters
  2. Workspace collaboration — shared saved prompts and role-based access on the portal
  3. Self-hosted NLP engine for organizations that cannot send anything to a cloud endpoint

If you want to shape what we build, use the tool and tell us where it falls short.

Contact