A context window the size of your database.
Polygres turns Postgres into working memory for AI agents. Rows, relationships, and embeddings, retrievable in one hybrid query. Stop stuffing prompts. Start querying.
Product HuntThree kinds of memory. One query.
Rows, relationships, and embeddings live in the same Postgres instance. Context assembly is a database query, not a pipeline.
Structured records
Your tables are already the ground truth. Rows, filters, and aggregates stream straight into context. No ETL. No sync jobs.
Connected relationships
pgGraph walks the connections that matter. The customer's orders. The ticket's history. Your agents reason over relationships, not just rows.
Semantic recall
HNSW vector search, native to your database. The relevant slice of everything you know, ranked and filtered in the same query.
Context assembly is one call.
Say what the agent needs. Semantic match, graph hops, scalar filters. Polygres hands back a ranked, token-ready context block. No vector store. No sync layer. No glue code.
Read the documentation# everything the agent needs, in one query
context = polygres.retrieve(
query="why did order #8231 fail?",
graph_hops=2, # orders → payments → disputes
filters={"status": "failed"},
limit_tokens=8_000,
)
agent.run(context) # grounded in your databaseStop rationing tokens.
Models cap out at a few million tokens. Your database doesn't. Retrieval turns every row into working memory, fetched the moment your agent needs it.
Give your agents the whole database.
Deploy a managed instance in seconds. Or point the open source SDK at your existing schema.
Start Building with Polygres
Get full access to relational table definitions, pgGraph traversals, and pgVector similarity search instantly.