AI Development Analysis
Libre WebUI includes repository scripts that can use local Ollama to summarize changes, generate changelog drafts, and inspect project health. These tools are for maintainers and run from the local checkout.
Scripts
| Command | Purpose |
|---|---|
npm run changelog | Generate a conventional changelog preview |
npm run changelog:ai | Generate an AI-assisted changelog |
npm run changelog:ai:summary | Generate a concise AI summary |
npm run changelog:ai:impact | Generate technical impact notes |
npm run analyze | Run project analysis with optional AI insight |
npm run analyze:quick | Run analysis without AI |
npm run analyze:metrics | Print metrics only |
Requirements
- Ollama running locally or reachable through
OLLAMA_BASE_URL - A local model installed
- Git history available in the repository
Good local model choices:
ollama pull gemma3:4b
ollama pull qwen3:8b
The scripts expose model overrides:
CHANGELOG_AI_MODEL=gemma3:4b npm run changelog:ai
ANALYSIS_AI_MODEL=qwen3:8b npm run analyze
If Ollama is unavailable, the changelog scripts fall back to non-AI generation where supported.
Environment Variables
| Variable | Purpose |
|---|---|
OLLAMA_BASE_URL | Ollama server URL, default http://localhost:11434 |
CHANGELOG_AI_MODEL | Model used by AI changelog scripts |
ANALYSIS_AI_MODEL | Model used by development analysis |
Good Release Notes
Use AI output as a draft, then edit it. A good changelog entry should:
- Describe user-visible changes first.
- Group fixes, features, docs, and internal work.
- Avoid implementation noise unless it affects users or operators.
- Mention breaking changes clearly.
- Link to commits or pull requests when useful.