A lightweight, stateless database for agent memory
Details
- External ID
- 49450816
- Source
- HN
- Company
- —
- Product
- A lightweight, stateless database for agent memory
- Website domain
- polign.com
- Launched
- Aug. 26, 2026
- Cohort
- —
- Upvotes
- 36
- Upvotes percentile
- 0.8245967741935484
- Tags
- —
- Fetched at
- Sept. 10, 2026, 5:32 a.m.
- Updated at
- Sept. 10, 2026, 5:32 a.m.
Description
I've been working on Polign and built a small prototype around something I've been thinking about with agent memory. I have built a lightweight/stateless vector db + BM25 search which works really well with typed facts and structured queries.It uses your own S3, or GCS bucket as primary storage, and restarting a node is fairly quick.Demo + writeup: https://polign.com/blog-edge-agent-memoryLive search demo: https://demo.polign.comDocs: https://polign.com
Enrichment
- Theme
- database infrastructure and developer tools
- Vertical
- Horizontal
- Function
- Data infrastructure
- Audience
- Developer
- AI stance
- AI feature
- Project type
- Commercial product
- Normalized one-liner
- stateless database for agent memory
- Manually corrected
- False
Could you build this?
Partial The user-facing API is straightforward, but designing a reliable stateless vector and BM25 index that operates efficiently directly against object storage requires deep database systems knowledge.
What it would actually take: Requires writing an embedded search engine in Rust or Go that formats data into immutable columnar and inverted-index files (similar to LanceDB or Tantivy) stored on S3/GCS. The challenging components include implementing byte-range queries over object storage, fast client-side HNSW/IVF vector indexing, and deterministic fact-conflict resolution semantics without persistent server state.
Discussion
14 comments analyzed.
Competitors mentioned: Vector databases (hosted solutions like Pinecone, Weaviate), Litestream with SQLite, Claude Code with markdown file indexing, LLM-generated custom memory systems
Concerns raised: Closed source approach is a non-starter for adoption, No clear differentiation from existing vector DB solutions, Lack of built-in observability and mutation capabilities, Keyword search limitations and ambiguity in retrieval, Cold query latency tradeoff
Feature requests: Observability and user-friendly mutation/modification interface, Open-source the core or release parts of it, Repo-aware agent memory for coding agents, Support for multiple memory types (working, procedural, semantic)
Competitors
Other products that read as similar to this one — 86 launches clear the similarity bar, closest 8 shown.
Attention rank: #20 of 87 (itself plus its competitors, highest first — normalized so YC and Product Hunt are compared fairly).
Launched 281 days after the earliest competitor.
- Remembrane · hn · 2026-08-07 · 13 upvotes · similarity 0.53
- Polign: Corrective Recall for Agents · ph · 2026-09-23 · 2 upvotes · similarity 0.47
- ClawMem · hn · 2026-03-22 · 5 upvotes · similarity 0.45
- YourMemory, agentic memory is a pruning problem, not a hoarding problem · hn · 2026-06-07 · 19 upvotes · similarity 0.44
- PMB · hn · 2026-06-22 · 7 upvotes · similarity 0.43
- We made a back end agnostic SDK for Agent Memory · hn · 2026-05-15 · 5 upvotes · similarity 0.42
- MouseBase · ph · 2026-09-08 · 1 upvotes · similarity 0.42
- Memori · ph · 2026-05-28 · 168 upvotes · similarity 0.42
Other launches for this product
- No other launches for this product.
Same idea, different domain
Nobody's really built a data infrastructure tool for Media & entertainment yet.