Metadata-Version: 2.4
Name: hydrogen-ai-sqlitevec
Version: 0.1.0
Summary: sqlite-vec vector store for hydrogen-ai
License: MIT
Author: phillychi3
Author-email: phillychi3@gmail.com
Requires-Python: >=3.9
Classifier: License :: OSI Approved :: MIT License
Classifier: Programming Language :: Python :: 3
Classifier: Programming Language :: Python :: 3.9
Classifier: Programming Language :: Python :: 3.10
Classifier: Programming Language :: Python :: 3.11
Classifier: Programming Language :: Python :: 3.12
Classifier: Programming Language :: Python :: 3.13
Classifier: Programming Language :: Python :: 3.14
Requires-Dist: hydrogen-ai (>=0.2.0,<0.3.0)
Requires-Dist: sqlite-vec (>=0.1.6,<0.2.0)
Description-Content-Type: text/markdown

# hydrogen-ai-sqlitevec

Local vector store for [hydrogen-ai](../../README.md) in a single SQLite file, using
[sqlite-vec](https://github.com/asg017/sqlite-vec). No server needed.

```bash
pip install hydrogen-ai-sqlitevec
```

```python
from hai.vectorstores import VectorStoreRetriever, VectorStoreWriter
from hai.vectorstores.sqlitevec import SqliteVecStore

store = SqliteVecStore(path="hai.db", collection_name="docs")

writer = VectorStoreWriter(store=store)
retriever = VectorStoreRetriever(store=store, top_k=5)
```

The tables are created on first use. Their vector size is inferred from the first
embedding unless `embedding_dim` is given.

sqlite-vec is a SQLite extension, so Python's `sqlite3` must be able to load
extensions. Builds from python.org, Homebrew, uv and pyenv can; the Python that
ships with macOS cannot.

