I’ve been working on a tool to query data using AI, but while testing it, I ran into something interesting.
The problem isn't writing queries.
Even when the AI generates the correct query, the result often comes back like:
userId
customerId
orderId
Technically correct - but not very useful.
So instead of focusing only on query generation, I started focusing on making results more readable.
QueryCanvas connects to databases, files, and APIs, lets you ask questions in plain English, and tries to return answers that actually make sense (names, emails, etc.), not just raw IDs.
You can also use a visual query builder, build charts, and create dashboards on top of it.
Still early and improving - especially around complex schemas - but would really appreciate feedback from anyone working with data.
Happy to share access if anyone wants to try it.
Would also love to hear how you're currently querying your data.