Choose the source connection.
Use the direct Postgres connection for your Supabase project, with credentials and network access suitable for replication. A transaction pooler is not the replication endpoint.
Loading page
SUPABASE + POLYGRES
Keep your application on Supabase. Connect your records to a unified retrieval layer for AI Search and graph retrieval.
YOUR DATABASE. ONE RETRIEVAL LAYER.
Your data stays authoritative. Polygres supplies the retrieval interface your agents use.
Use the direct Postgres connection for your Supabase project, with credentials and network access suitable for replication. A transaction pooler is not the replication endpoint.
Check your source’s logical replication settings and the permissions required for publication and replication. Select tables with a stable primary key or suitable unique identity. Polygres synchronizes selected tables in the public schema. Check the provider-specific guide for the exact setup and permissions.
Connect your source and choose the records your agents need. Polygres brings graph, semantic, and text retrieval behind one interface, so your application can ask for context without assembling each retrieval service.
Configure AI Context collections and graph relationships for your application. Use AI Search or graph retrieval through the Python SDK or Runtime API. Your source database remains the place for writes and schema changes.
Keep your application, business logic, and writes on your existing database. Polygres provides the retrieval layer between those records and the agents that use them.
Connected projects currently keep selected data synchronized for retrieval. The source guide covers setup and freshness. Choose project capacity for your data and indexes as you build.
Estimate your project capacity ↗LET’S BUILD
Start with your Postgres. Explore what Polygres can do.