This roadmap tracks the next steps after the Broadside-AI v1 cleanup pass.
The project now has:
- a CLI-first scatter/gather runtime
- stdout-first default behavior for automation
- stable
--json-outputpayloads - task validation from the main CLI
- structured-output parsing through gather and synthesis
weighted_mergefor structured tasks- early-stop controls for cost and latency
- tests, benchmarks, CI, and release/publish workflows
- publish the first package to PyPI
- cut the initial GitHub release notes
- verify Trusted Publishing end to end
- confirm README rendering and metadata on PyPI
- add dedicated backend setup docs
- document Ollama local vs cloud troubleshooting
- document OpenAI-compatible provider configuration
- add a "when not to use Broadside-AI" guide
- add more subprocess and CI examples
- document JSON payload versioning expectations
- add examples for structured extraction and ticket triage
- add benchmark recipes for comparing backends
- add backend contract tests with fake clients
- add more CLI smoke tests in CI
- add release smoke tests for example workflows
- add benchmark-result sanity checks
add issue templates and a PR template(done)- label good first issues
- document benchmark contribution expectations
- richer conflict detection on structured outputs
- model-diversity scatter strategies
- optional result schema versioning for JSON mode
- hosted documentation site
These remain out of scope unless the product identity changes:
- inter-agent messaging
- workflow DAG orchestration
- long-running autonomous agents
- crew or role hierarchies
- hidden persistent memory across runs
The repository markdown files are part of the product surface. If behavior changes, update the docs in the same PR.
Last updated: 2026-04-05