A lightweight API wrapper + AI-powered scanner that connects to any database and surfaces bad data, bias, outliers, and inconsistencies in seconds.
12 automated checks covering every dimension of data quality, powered by AI analysis.
Per-field null percentage with severity scoring. Catches empty strings, N/A values, and hidden missing data.
Exact record deduplication plus key-field uniqueness checks for IDs, emails, and other constrained fields.
Detects mixed types, numeric strings, and mismatched formats across all fields automatically.
IQR-based outlier detection flags extreme values that may indicate data entry errors or corruption.
Identifies gender skew, category imbalance, and collection bias patterns across demographic fields.
Claude streams a full executive summary with risk assessment, field recommendations, and remediation steps.
Connect to any REST API database or load demo data to see the analyzer in action.
Five built-in adapters. Drop in your connection string and you're live.
Every table in your database exposed as a fully filterable, paginated JSON API.
Zero database required to start — SQLite auto-creates a demo table on first run.
Download the project files and install dependencies. Node.js 18+ required.
Edit config.json — set type to your database and fill in credentials. Leave as-is to use the built-in SQLite demo.
The API wrapper starts on port 3000. Then paste the endpoint URL into the analyzer above.
Manage features, access, and display settings for DataPulse.