Analyze your data with AI. Keep it private.
Ask Claude and ChatGPT questions about your large CSV files without uploading anything to the cloud.
AI is great at analyzing data — until you have to upload it.
Business owners want to explore their numbers with AI. But pasting CSVs into ChatGPT burns API credits, hits context limits, and quietly sends customer records to a third party. Meanwhile, tech teams need clear metric definitions to ship dashboards — and email threads aren't enough.
Three steps from CSV to insight
Import
Drag CSV files into DashDraft. Data stays on your computer.
Ask
Connect Claude or ChatGPT. Ask questions. Get answers without uploading raw data.
Share
Hand off chat history and SQL logic so your tech team can build production dashboards.
Built for business owners, loved by tech teams
Data privacy by default
Your CSV files never leave your machine. AI only sees query results — never raw rows. Hash sensitive columns automatically.
Zero API credit waste
No bulk CSV uploads to OpenAI or Anthropic. Only aggregated results flow to the model. Save money on API costs.
POC your metrics
Test metric definitions before building dashboards. Validate logic with stakeholders. Reduce back-and-forth with tech teams.
SQL history for handoff
Every query is logged automatically. Share exact SQL with developers. Keep a clear audit trail of metric definitions.
Works with Claude & ChatGPT
Direct stdio connection for Claude Desktop. Secure ngrok tunnel for ChatGPT. Same data, two AI options.
Offline-first & fast
SQL engine for fast local queries. No internet needed after install. Instant workspace switching.
Real teams, real questions
Understanding customer churn
A PM has a CSV of sign-ups and cancellations and wants churn by region. Instead of uploading the customer database to ChatGPT, they run the query locally with DashDraft, get the rate (West: 12%, East: 8%), and share the SQL with their data team.
Monthly revenue analysis
A business owner imports order data, asks Claude to calculate revenue by month and product category, then saves the SQL logic to automate the report in their BI tool.
Building metric definitions
An operations lead needs to define 'active users.' They experiment with SQL (last 7 vs. last 30 days), validate the numbers with their team, then hand off the final query.
One command. No dependencies.
DashDraft ships as a single binary. No Node, no Python, no Docker. Installs in under 2 minutes and runs entirely on your computer.
Read full installation guidecurl -fsSL https://dashdraft.app/get/install.sh | bashirm https://dashdraft.app/get/install.ps1 | iexWatch DashDraft at work
Start analyzing your data privately
Free, local, and ready in two minutes.