VQV Signal

AI AT WORK SOURCE-BACKED TECHNICAL

Anathem offers local memory for coding agents with no LLM, daemon, or HTTP server

Anathem is a Show HN project described as local memory for coding agents such as Claude Code and opencode. It is a small binary that stores memories in a SQLite file, with no daemon and no HTTP server. According to the source, saving works without a model or an API key, and the agent decides what to save while Anathem stores and indexes those memories. Search uses BM25 over FTS5.

Source: Hacker News Newest · gitlab.com Published 2026-11-11T18:10:46+00:00 Detected 2026-11-11T22:19:29+00:00
View original source

Anathem is a Show HN project described as local memory for coding agents such as Claude Code and opencode. It is a small binary that stores memories in a SQLite file, with no daemon and no HTTP server. According to the source, saving works without a model or an API key, and the agent decides what to save while Anathem stores and indexes those memories. Search uses BM25 over FTS5.

AI-assisted summary based on the listed source.

hi, anathem is a local memory for agents (like Claude Code and opencode): a small binary and a SQLite file, no daemon and no HTTP server. saving works without a model or an API key. the agent decides what to save. anathem just stores those memories and makes them searchable. search is BM25 over FTS5.

Developers who run coding agents locally may find this useful if they want persistent memory without running extra services or supplying API keys. The design keeps storage simple, since the memory lives in a single SQLite file that the agent can query. Users should check the project's documentation and test it on their own setup to see how well search works in practice.

Teams using AI at work may want to compare this against current productivity and review workflows.

Signal Strength 95% Technical label SOURCE-BACKED Public Interest 52 Category AI AT WORK Reader Depth TECHNICAL Event context 1 source

Signal Strength reflects source quality, relevance, freshness and evidence. Public Interest helps organize discovery; it is not proof of truth.

Public Interest components
Recognizable Entity Score 51 Practical Impact Score 28 Novelty Interest Score 94 Consequence Score 46 Curiosity Score 16 Shareability Score 61

Claude gets a source-backed update

Claude has a source-backed update with coverage spanning for developers.

1 source 1 angle FOR DEVELOPERS

VQV surfaced this signal because it is recent, relevant to AI Coding Tools, connected to Hacker News Newest.