Polygres
Hybrid Postgres retrieval and internal search for AI Agents
About Polygres
Polygres turns an existing Postgres into an internal search layer for AI Agents, unifying retrieval of rows, relationships, and embeddings. It queries graph traversal and ten search methods in parallel, then fuses them into ranked context ready to provide to a model. It supports a Polygres-managed database or connections to Supabase, Neon, and other Postgres providers that supply a connection string. pgGraph and pgVector cover retrieval paths from dense vectors to graph relationships.
Best for
Best use cases
Retrieve business data for AI Agents
Search rows and relationships in Postgres
Build personal projects or early-stage Agent applications
Scale Agents in production
Features
Hybrid queries
Run graph traversal and ten search methods in parallel, then fuse them into ranked context.
Postgres connections
Use a managed database or connect Supabase, Neon, and other Postgres providers that supply a connection string.
Open-source retrieval engines
pgGraph and pgVector cover retrieval paths from dense vectors to graph relationships.
Self-hosting plan
The Self-Hosted plan runs pgGraph and pgVector on your own Postgres instance.
Pros
- Run open-source engines on your own Postgres
- Connect to Supabase, Neon, and other providers
- Fuse multiple retrieval results into ranked context
Pricing Plans
Self-Hosted
- Open-source pgGraph & pgVector
- Run on any Postgres instance
- Community Support
Launch
- Managed cloud database hosting
- Support: Email & Community Discord
Scale
- Automated vertical & horizontal scaling
- Support: Email & Community Discord
- Private Discord/Slack Channel
- Support Hours: 8:00 AM - 5:00 PM
* Pricing information is for reference only. Please refer to the latest announcement on Polygres's official website.