Rekal
Long-term memory for LLMs in a single SQLite file
Details
- External ID
- 47744683
- Source
- HN
- Company
- —
- Product
- Rekal
- Website domain
- github.com
- Launched
- April 12, 2026
- Cohort
- —
- Upvotes
- 9
- Upvotes percentile
- 0.5501285347043702
- Tags
- —
- Fetched at
- Sept. 7, 2026, 9:26 p.m.
- Updated at
- Sept. 7, 2026, 9:26 p.m.
Description
I got tired of repeating myself to my LLM every session. rekal is an MCP server that stores memories in SQLite and retrieves them with hybrid search (BM25 + vectors + recency decay). One file, local embeddings, no API keys.
Enrichment
- Theme
- database infrastructure and developer tools
- Vertical
- Horizontal
- Function
- Model & infra
- Audience
- Developer
- AI stance
- AI-native
- Project type
- Commercial product
- Normalized one-liner
- persistent memory for llms
- Manually corrected
- False
Could you build this?
Yes Rekal is an MCP server written on top of SQLite, sqlite-vec / FTS5 for hybrid BM25 and vector search, and an embedded local model (e.g. fastembed/onnx). This is a well-scoped weekend developer tool that an AI assistant can easily generate.
Discussion
10 comments analyzed.
Concerns raised: Python 3.14 requirement may limit adoption, Effectiveness unclear - requires significant intervention to keep agents using guardrails, Fixed weights (0.4/0.4/0.2) may not suit different use cases
Feature requests: Make hybrid search weights configurable, Add config file support (pyproject.toml or .rekal/config.yml), Different decay rates for different memory types
Competitors
Other products that read as similar to this one — 31 launches clear the similarity bar, closest 8 shown.
Attention rank: #16 of 32 (itself plus its competitors, highest first — normalized so YC and Product Hunt are compared fairly).
Launched 136 days after the earliest competitor.
- Remembrane · hn · 2026-08-07 · 13 upvotes · similarity 0.55
- Llm.sql · hn · 2026-04-24 · 8 upvotes · similarity 0.50
- A local-first memory store for LLM agents (SQLite) · hn · 2025-12-14 · 48 upvotes · similarity 0.42
- Continual Learning with .md · hn · 2026-04-13 · 34 upvotes · similarity 0.40
- Rekall · hn · 2026-08-12 · 5 upvotes · similarity 0.37
- Engrim · hn · 2026-09-07 · 92 upvotes · similarity 0.37
- Lightweight Task queue on Erlang/OTP, SQLite-backed, no overengineering · hn · 2026-06-10 · 75 upvotes · similarity 0.36
- UseSQLite · ph · 2026-09-16 · 2 upvotes · similarity 0.36
Other launches for this product
- No other launches for this product.
Same idea, different domain
Nobody's really built a model & infra tool for Fintech yet.