Add SQLite-backed registry with JSON migration support

.github/workflows/publish.yml

@@ -8,12 +8,13 @@ on:
 jobs:
   publish:
     runs-on: ubuntu-latest
+    environment: pypi
     permissions:
       id-token: write
     steps:
-      - uses: actions/checkout@v4
+      - uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332  # v4.2.2
 
-      - uses: actions/setup-python@v5
+      - uses: actions/setup-python@f677139bbe7f9c59b41e40162b753c062f5d49a3  # v5.2.0
         with:
           python-version: "3.12"
 
@@ -24,4 +25,4 @@ jobs:
         run: python -m build
 
       - name: Publish to PyPI
-        uses: pypa/gh-action-pypi-publish@release/v1
+        uses: pypa/gh-action-pypi-publish@fb13cb306901256ace3dab689990e13a5550ffaa  # release/v1.11.0

README.md

@@ -147,8 +147,20 @@ Settings are initialized by `openkb init`, and stored in `.openkb/config.yaml`:
 model: gpt-5.4                   # LLM model (any LiteLLM-supported provider)
 language: en                     # Wiki output language
 pageindex_threshold: 20          # PDF pages threshold for PageIndex
+storage_backend: sqlite          # Storage backend: sqlite (default) or json
 ```
 
+### Storage Backend
+
+OpenKB supports two storage backends for the file hash registry:
+
+| Backend | Description | Use Case |
+|---------|-------------|----------|
+| `sqlite` | SQLite database (default) | Better concurrency, scalability, recommended for production |
+| `json` | JSON file | Simple, human-readable, for small installations |
+
+Migration from JSON to SQLite happens automatically when you switch to `sqlite` backend and a `hashes.json` file exists. The JSON file is preserved but no longer used.
+
 Model names use `provider/model` LiteLLM [format](https://docs.litellm.ai/docs/providers) (OpenAI models can omit the prefix):
 
 | Provider | Model example |