Metadata-Version: 2.3
Name: vexicon
Version: 0.5.0
Summary: Hybrid vector + keyword MCP server powered by chroma and sqlite
Author: estasney
Author-email: estasney <estasney@users.noreply.github.com>
Requires-Dist: aiosqlite>=0.22.1,<0.23.0
Requires-Dist: alembic>=1.19.1,<2.0.0
Requires-Dist: chromadb>=1.5.9,<2.0.0
Requires-Dist: fastmcp>=4.0.10,<5.0.0
Requires-Dist: huggingface-hub>=1.28.0,<2.0.0
Requires-Dist: numpy>=2.5.2,<3.0.0
Requires-Dist: pydantic>=2.13.4,<3.0.0
Requires-Dist: pydantic-settings>=2.15.0,<3.0.0
Requires-Dist: sentence-transformers>=6.0.0,<7.0.0
Requires-Dist: sqlalchemy>=2.0.52,<3.0.0
Requires-Python: >=3.12
Description-Content-Type: text/markdown

# vexicon

## About

Knowledge store MCP Server.

Provides an RRF implementation using SQLite. Tunable via settings.

Introduces some opinionated defaults


Notes and reference material live in
named spaces backed by Chroma collections with a SQLite FTS5 keyword index;
searches fuse the vector and keyword rankings by reciprocal rank fusion.

Run the server over stdio with `uv run vexicon`.

Settings are read from `VEXICON_*` environment variables; see
`src/vexicon/settings.py`. Data is stored under `~/.vexicon` unless
`VEXICON_PERSISTENT_PATH` and `VEXICON_INDEX_DB_PATH` say otherwise.

## Settings

| Variable                  | Default                | Meaning                                                                   |
|---------------------------|------------------------|---------------------------------------------------------------------------|
| `VEXICON_PERSISTENT_PATH` | `~/.vexicon/chroma`    | Directory where Chroma stores its database.                               |
| `VEXICON_INDEX_DB_PATH`   | `~/.vexicon/hybrid.db` | SQLite file holding the keyword index.                                    |
| `VEXICON_VECTOR_WEIGHT`   | `1.0`                  | Weight of the vector ranking in fusion.                                   |
| `VEXICON_KEYWORD_WEIGHT`  | `1.0`                  | Weight of the keyword ranking in fusion.                                  |
| `VEXICON_RRF_RANK_OFFSET` | `60`                   | Rank offset in reciprocal rank fusion.                                    |
| `VEXICON_DEVICE`          | `auto`                 | Device that runs embedding models: `auto`, `cpu`, or `cuda`.              |
| `VEXICON_IDLE_SECONDS`    | `300`                  | Seconds without activity before Chroma and embedding models are unloaded. |

## Compared with chroma-mcp

Compared against chroma-mcp 0.2.6 and chromadb 1.5.9.

### Hybrid search without Chroma Cloud

- vexicon combines vector and BM25 keyword rankings with reciprocal rank
  fusion, using a SQLite FTS5 index beside each local Chroma collection.

### Token usage

- vexicon's tool definitions take about 30% fewer tokens than chroma-mcp's.
- `search` returns at most `limit` entries in total, while chroma-mcp returns
  `n_results` per query, so its response grows with every query a model adds.
- Results always leave out embedding vectors, which add tokens without giving
  a model anything it can use.

### Embedding models

- Each space can use any sentence-transformers model, and tools let a model
  find and download one from Hugging Face.
- Cached models load without network calls.
- Models trained with separate query and document prompts get the matching
  prompt for searches and for stored entries.
- Spaces always have their embedding model's max token size available in
  metadata for sizing entries before they are added.

### Memory

- Embedding models unload from memory after `VEXICON_IDLE_SECONDS` without
  use.

### Defaults

- Storage persists under `~/.vexicon` by default.
- Entry IDs are kebab-case mnemonics.
- Duplicate entry IDs raise an error.
- Spaces are published as MCP resources.
- Chroma telemetry is off.

vexicon keeps its own fields in Chroma metadata:

- Each entry's metadata holds `created_at` in epoch seconds for recency
  filters in `where`.
- Each space's metadata holds its `readme`, `embedding_repo_id`, and
  `embedding_max_tokens`.
- Callers cannot set those three keys through a space's `metadata` argument.
- Tool results show these fields apart from the caller's own metadata.
