Show HN: CamelAI – Perplexity for Your Data
3 points by vercantez 6 months ago | 1 commentWhy we built this: - Business teams waste hours waiting for data analysts to build dashboards - Non-technical founders can't get quick insights from their own data - Writing SQL queries is a barrier to data-driven decisions
What makes it different: - Uses DuckDB to query terabytes in seconds (built-in, no setup needed) - Supports advanced analytics (cohort analysis, statistical computations) not just basic SQL - Multi-turn conversations: AI explores your schema, analyzes results, and creates interactive charts - Explains its reasoning and suggests follow-up analyses
Tech details: - Query engine: DuckDB for lightning-fast analytics - AI: Claude Sonnet for chat agent, o1 for report generation - Native connectors: Postgres, MySQL, Supabase, BigQuery (beta) (more to come) - Also supports: CSV, Excel (beta)
Video demo: https://camelai.com/demo/ Try it free (10 messages/week): https://app.camelai.com
Example questions you can ask: - "What's driving our recent user churn?" - "Show me conversion rates by marketing channel" - "Which features correlate with user retention?"
We're a team of ex-Apple/Google engineers who believe data analysis shouldn't require a CS degree. Would love HN's feedback, especially on our query generation and accuracy.
- gpsmsn 6 months agoAwesome demo! I piloted a very similar product at my then 15 person company so hopefully you find this feedback useful:
The product seems great if you're familiar with the data and schemas, but for a product like this to appeal to teams and have impact for anyone on a team there needs to be (1) a way to provide system prompts and add business/product context to the underlying schema (2) a way to store historical queries that can be used as a starting point for other users, much like saving a query, model, or dashboard in a BI tool
I hope to see more news about CamelAI in the future. Good luck!