Squish scans your databases and maps every relationship automatically, including connections across different services.
Why Squish
Data relationships exist in people's heads.
Engineers leave. Slack messages get lost. ERDs go stale. Squish builds a live map that stays current.
Cross-Database Discovery
Find relationships across SQL and NoSQL databases. PostgreSQL, Snowflake, MongoDB, DynamoDB -- Squish discovers connections between them all.
No More Slack Messages
Stop asking "which table has customer IDs?" and waiting days for answers. Get a live, searchable map of how your data connects.
Complements Your dbt Workflow
Auto-generate the relationship documentation that dbt requires you to write by hand. Export to MetricFlow, Snowflake Cortex, or Databricks.
Context for AI Tools
Give Snowflake Intelligence and Databricks Genie the semantic context they need. Better context means more accurate AI-generated queries.
Statistical Analysis, Not Guesswork
We analyze actual data patterns, not just column names. Configurable confidence thresholds control what gets surfaced.
Schema Change Detection
Get notified when schemas change and relationships break. Your ERD stays accurate even as your data structure evolves.
How it works
Four steps to mapped relationships
Connect your databases. Get a live relationship map.
Connect your databases
Add PostgreSQL, Snowflake, BigQuery, or MySQL connections. We use read-only access and only sample schema metadata.
We analyze patterns and data
Squish extracts foreign keys, matches naming patterns, and runs statistical analysis on sample data to find hidden relationships.
Review with confidence scores
Every discovered relationship comes with an explainable confidence score. Mark false positives to improve accuracy over time.
Export or connect your tools
Generate semantic layer definitions for dbt, Snowflake Cortex, or Databricks. Or expose context to AI tools via MCP.
Products
From discovery to AI-ready semantic layers
Discover, document, and generate semantic context for AI tools.
Relationship Discovery
Discover
Find relationships that exist only as tribal knowledge. Cross-database discovery, statistical analysis, and explainable confidence scores.
- Cross-database relationship detection
- Statistical analysis, not just foreign keys
- Configurable confidence thresholds
- Continuous schema monitoring
- Interactive ERD visualization
Data Documentation
Catalog
A catalog people actually use. Auto-populated from discovery, with AI-powered search, quality scoring, and team collaboration.
- AI-powered hybrid search with vector embeddings
- Trust scores across 7 quality dimensions
- AI-generated table and column descriptions
- Classification and PII detection
- Business glossary with approval workflows
AI-Ready Context
Semantic Layer
Auto-generate semantic models from discovered relationships. Export to dbt, Snowflake Cortex, Databricks, and more.
- Auto-generated from discovery
- Export to 9 platforms including dbt and Snowflake
- LookML, Cube.dev, Power BI, ThoughtSpot
- Open Semantic Interchange standard
Free tier available. No credit card required.
Integrations
24+ Databases Supported
Connect any SQL-compatible database. More added every month.
Security
Your data stays in your database
Squish reads only schema metadata. Your actual data never leaves your infrastructure.
View security details->Read-Only Access
Squish only reads database metadata. We never modify your data or access actual row content.
AES-256 Encryption
Database credentials encrypted at rest with AES-256. All connections use TLS in transit.
Role-Based Access
Organization-level isolation with role-based access control for team members.
Metadata Only
We query information_schema for table names, column names, and data types. Never your actual data.
Stop reverse-engineering your data
Map your data relationships automatically so your team can understand the system without archaeology.