Lab notes.
What we shipped, what broke, what we measured. Pretensor is open-source and in beta — this is where we explain how it works under the hood.
- mcp · dbt · lineage · schema-graph · data-engineering
dbt lineage and schema graphs: the missing MCP integration
dbt knows your data transformations. Pretensor knows your schema structure. Together, they give AI agents the full picture of how data flows from source to dashboard.
- mcp · comparison · database · tools
MCP database tools: what exists and what's missing
A survey of MCP servers for databases — raw SQL executors, metadata readers, and the schema-intelligent tools that actually understand your tables.
- mcp · cross-database · postgres · snowflake · bigquery · schema-graph
One schema graph for Postgres, Snowflake, and BigQuery
Most AI agents are single-database. Pretensor builds a unified schema graph across Postgres, Snowflake, and BigQuery — same tools, same interface, cross-DB join paths included.
- mcp · claude · cursor · postgres · tutorial
How to connect Claude or Cursor to your database with schema context
A practical guide to setting up Pretensor as an MCP server so your AI agent understands your database schema — not just runs SQL against it.
- mcp · schema · impact-analysis · change-detection · migrations
What happens when you change your database schema? AI agents should know.
Schema changes break queries. Pretensor's impact analysis and change detection tools give AI agents the awareness they need to avoid producing wrong answers.
- mcp · schema · database · introspection
Schema introspection for MCP: why your AI agent needs more than SQL
Raw SQL access gives AI agents a firehose. Schema introspection gives them a map. Here's why the difference matters for MCP servers.
- mcp · database · schema · knowledge-graph · ai-agents
Why your AI agent needs a schema graph, not just SQL access
Most MCP database servers let agents run raw SQL. That works at 30 tables and breaks at 200. A schema graph gives the agent FK-backed join paths, impact analysis, and table context — structural knowledge, not just query access.