One API to search, scrape, crawl, map and monitor the web, returning clean structured text that AI agents and RAG pipelines can use directly.
Data & AnalyticsFreemium
Polygres
by Evokoa
Added to Onei
0 bookmarks
Summary
Turn a Postgres database into working memory for AI agents — relational rows, graph traversal and vector search fused into one ranked, token-budgeted context block.
Screenshots
Description
Retrieval stacks tend to fragment: a relational database for the records, a vector store for embeddings, a graph database for relationships, and sync pipelines holding the three in uneasy agreement. Polygres collapses that into one Postgres 17 instance and one hybrid API.
One query, several retrieval strategies
A Polygres query fans out across graph traversal and ten search methods simultaneously, then fuses the results into a single ranked block sized to the agent's token budget. The point is that structured rows, connected relationships, semantic matches and full-text hits arrive together and already ordered, rather than as three result sets the application has to reconcile and truncate itself. Whether the underlying table holds ten rows or ten million, the prompt keeps the same shape.
The two engines underneath
Both are open-source Postgres extensions maintained by Evokoa:
- pgGraph models relationships natively — traverse connections, discover neighbours and map entity graphs without hand-writing recursive SQL joins.
- pgContext stores embeddings and runs similarity search, with HNSW indexing for fast lookups at scale and scalar filtering alongside.
Because it is ordinary Postgres, ACID guarantees and the standard tooling ecosystem still apply — this is not a separate datastore pretending to be a database.
Who it is for
Teams building agents and retrieval-augmented applications that need to reason over structured relationships as well as unstructured text, and who have watched a vector-only or graph-only setup miss context that was plainly present in the data.
Pricing
Self-hosting is free forever on any Postgres instance with the open-source extensions and community support. The managed Launch plan is listed at $16/month and currently free; Scale ($256/month minimum usage, business-hours support and a private channel) and Enterprise ($4,096/month minimum, dedicated infrastructure and 24/7 SLA support) are announced as coming soon. Managed cloud deployment scales through Kubernetes and Docker.
Built by Evokoa.
Reviews
Similar App Suggestions
An OpenTelemetry-native control plane for AI agents that enforces policy mid-run instead of paging you after the incident, with cost attribution and EU AI Act evidence built in.
Data & AnalyticsFreemium
A vectorless RAG engine that builds a tree index of long PDFs and lets an LLM reason its way to the right section, with page-level citations.
Data & AnalyticsFreemium
Agentic data platform that turns business questions into defensible, auditable narratives — with sources, assumptions and lineage attached.
Data & AnalyticsSubscription
Reads your production chat and voice agent conversations to surface the silent failures, frustration loops and policy breaches that offline evals never catch.
Data & AnalyticsFreemium