neurographrag 2.0.2

Local GraphRAG memory for LLMs in a single SQLite file
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68

# neurographrag


> Persistent memory for 21 AI agents in a single 25 MB Rust binary

- Local GraphRAG memory combining FTS5, sqlite-vec KNN and entity graph
- Single SQLite file storage with deterministic JSON CLI output for orchestration
- Cross-platform native binary for Linux, macOS and Windows without external services


## Primary Documentation
### Canonical English sources for LLM ingestion
- [README](https://github.com/daniloaguiarbr/neurographrag/blob/main/README.md)
- [HOW_TO_USE](https://github.com/daniloaguiarbr/neurographrag/blob/main/docs/HOW_TO_USE.md)
- [CROSS_PLATFORM](https://github.com/daniloaguiarbr/neurographrag/blob/main/docs/CROSS_PLATFORM.md)
- [AGENTS persuasive guide](https://github.com/daniloaguiarbr/neurographrag/blob/main/docs/AGENTS.md)
- [AGENT_PROTOCOL imperative spec](https://github.com/daniloaguiarbr/neurographrag/blob/main/docs/AGENT_PROTOCOL.md)
- [COOKBOOK recipes](https://github.com/daniloaguiarbr/neurographrag/blob/main/docs/COOKBOOK.md)
- [INTEGRATIONS guide](https://github.com/daniloaguiarbr/neurographrag/blob/main/INTEGRATIONS.md)
- [CHANGELOG](https://github.com/daniloaguiarbr/neurographrag/blob/main/CHANGELOG.md)
- [CODE_OF_CONDUCT](https://github.com/daniloaguiarbr/neurographrag/blob/main/CODE_OF_CONDUCT.md)
- [SECURITY](https://github.com/daniloaguiarbr/neurographrag/blob/main/SECURITY.md)
- [CONTRIBUTING](https://github.com/daniloaguiarbr/neurographrag/blob/main/CONTRIBUTING.md)


## Core Commands
### Subcommands grouped by lifecycle
- `init` initializes the SQLite database and downloads the embedding model
- `remember` saves a memory with name, type, description, body and optional graph
- `recall` performs vector KNN similarity search over stored memories
- `hybrid-search` fuses FTS5 and vector KNN via Reciprocal Rank Fusion
- `read` `list` `forget` `rename` `edit` `history` `restore` manage memory lifecycle
- `health` `stats` `migrate` `vacuum` `optimize` `sync-safe-copy` manage the database
- `purge` `namespace-detect` handle maintenance and scope resolution


## Environment Variables
### Runtime configuration surface
- `NEUROGRAPHRAG_DB_PATH` absolute path to the SQLite database file
- `NEUROGRAPHRAG_CACHE_DIR` directory for embedding model cache files
- `NEUROGRAPHRAG_LOG_LEVEL` tracing filter level with values `error` through `trace`
- `NEUROGRAPHRAG_LANG` CLI output language with values `en` or `pt`
- `NEUROGRAPHRAG_NAMESPACE` namespace override bypassing auto detection


## Exit Codes
### Deterministic status for pipeline routing
- `0` success `1` validation `2` duplicate `3` conflict `4` not found
- `5` namespace `6` payload `10` sqlite `11` embedding `12` sqlite-vec
- `13` batch partial failure `14` filesystem `15` database busy `20` internal
- `73` memory guard `75` slots busy `77` low RAM


## Optional References
### Supplementary materials for deeper context
- [CLAUDE guidance](https://github.com/daniloaguiarbr/neurographrag/blob/main/docs/CLAUDE.md)
- [SKILL definitions](https://github.com/daniloaguiarbr/neurographrag/tree/main/skill/neurographrag-en)
- [crates.io package](https://crates.io/crates/neurographrag)
- [docs.rs API reference](https://docs.rs/neurographrag)


## Stable Facts
### Identity and version metadata
- Package name `neurographrag` published on crates.io under MIT OR Apache-2.0
- Current version 2.0.0 with MSRV Rust 1.88 declared in `Cargo.toml`
- Repository `https://github.com/daniloaguiarbr/neurographrag` with CI on push
- Embedding model `multilingual-e5-small` quantized via `fastembed` at approximately 750 MB RAM
- Storage layer `rusqlite` with bundled SQLite plus `sqlite-vec` extension and FTS5