Roadmap

What StockSage is

A local-first stock analysis tool you run on your own machine with your own model (Ollama or any supported cloud provider). Not a hosted service, not investment advice.

In scope — what we're investing in - Better analysis quality and explainability (stronger agent prompts, clearer reasoning in the final verdict). - Simple, reliable model-provider setup across local (Ollama) and cloud APIs. - Contributor-friendly docs and tests for the core pipeline.

Not planned: broker integrations, automated trading, hosted multi-user accounts, real-time market data SLAs.

What's next

Rough priority order. Open an issue if you'd like to pick one up.

1. Publish to PyPI so installing becomes pip install stocksage.
2. Automated releases — on merge to main, parse Conventional Commit titles to bump the version in pyproject.toml, regenerate CHANGELOG.md, and open a release PR; on a v*.*.* tag, build and publish the wheel to PyPI.
3. Troubleshooting guide — common setup issues (Ollama not reachable, provider rate limits, stale cached data).
4. More robust tests for the live SSE stream and higher coverage on the analysis pipeline.
5. Community surface — enable GitHub Discussions, seed a few good-first-issues, add searchable repo topics.
6. New analysis cards or agents — sector comparison, peer benchmarks, earnings-call sentiment. Ideas welcome in issues.
7. Concurrent analyses — support multiple tickers in parallel. Today a second request fails or reports "already in progress"; we want per-request isolation so different users (or tabs) can analyze independently.

License

MIT. See LICENSE.