Turn a CSV into a board-ready brief in seconds.

Your CSV is parsed in your browser. A compact summary — columns, types, summary stats, and a small sample of rows — is sent to Claude (an AI model from Anthropic). Claude returns insights and chart suggestions, and the page renders them. The raw file never leaves your machine.

The file itself never uploads anywhere. Only the digest goes to the AI, and that traffic is not used to train models. For sensitive data (PII, health, finance) I can set up a deployment that uses zero-retention endpoints or self-hosted models.

Three common patterns: a one-time setup tailored to your data, with a private URL your team can use; a recurring report connected to where your data already lives (Google Sheets, your CRM, a database) and posted to Slack or email; or embedded into a tool you already use, like a Notion page or a Slack bot. Tell me what you have and I will recommend.

A custom deployment with one of the integration patterns above runs roughly $2-5k for setup and $300-800/month to keep running. Specifics depend on data volume, how often it runs, and where it lives. The AI cost itself is small — usually pennies per report.

Two safeguards. The prompt forces the AI to cite specific columns and values from your data, which makes hallucinations rare and easy to spot. For high-stakes use cases — anything customer-facing or financial — I add a human-in-the-loop step where someone on your team reviews and approves the AI output before it ships. Your data stays on your side; I am not the reviewer.

Standard tabular data — clean header row, consistent columns — works great. Multi-tab Excel exports, reports with title rows and footers, or freeform notes need a preprocessing step. Send me a sample file and I will tell you straight whether it is in scope.

You would be among the first. The demo on this page shows the capability; deployments are bespoke. If that is a risk concern, I am open to a short paid pilot — small scope, refundable if it does not deliver.