Roadmap
Regressa 0.1 shipped the full pipeline: ingest, versioning, evals, alerts, dashboard, CI gate, governance and four SDKs. This is the plan for what comes next. Every item is a GitHub issue; ones tagged help wanted are open to collaborators and good first issue marks small, well-scoped starting points. Comment on an issue to claim it.
Milestones: 0.2 Hosted beta · 0.3 Quality workflows · 0.4 Enterprise
0.2 Hosted beta
Run Regressa for real users: deployment, billing enforcement, auth hardening, published SDKs. Target: 2026-10-15.
| # | Task | Labels |
|---|---|---|
| #1 | Hosted deployment: Timescale Cloud + Redis + container host, Terraform | help wanted |
| #2 | Enforce plan limits at ingest (traces/month) and surface usage in the UI | core team |
| #3 | Stripe: real price ids, customer portal, webhook coverage for upgrades and downgrades | help wanted |
| #4 | Password reset and email verification | help wanted |
| #5 | Publish SDKs: npm @regressa/node, PyPI regressa-sdk, Go module tag, Maven Central | core team |
| #6 | Integration tests against real Postgres/Redis in CI | help wanted |
| #7 | Playwright end-to-end tests for the dashboard | help wanted, good first issue |
| #8 | Self-host Inter and JetBrains Mono so next build works offline |
good first issue, help wanted |
| #9 | Rate limit the public REST API and gate endpoint per key | help wanted |
0.3 Quality workflows
Golden set management, playground compare, template registry, datasets export, OTel depth. Target: 2026-11-30.
| # | Task | Labels |
|---|---|---|
| #10 | Golden set management UI: add examples from traces, edit, re-embed | core team |
| #11 | Playground: compare against a stored version and save runs as traces | core team |
| #12 | Prompt template registry API: fetch the current template at runtime | core team |
| #13 | Export traces, eval results and golden sets as JSONL/CSV | good first issue, help wanted |
| #14 | Trace groups view for multi-call workflows | help wanted |
| #15 | Use continuous aggregates for 7d/30d dashboard ranges | help wanted |
| #16 | OpenTelemetry: gRPC receiver, protobuf OTLP, and tool-call spans | help wanted |
| #20 | Accessibility and responsive pass on the dashboard | good first issue, help wanted |
0.4 Enterprise
Slack app, SSO groups, RBAC, audit export, seasonality-aware anomalies, i18n. Target: 2027-01-31.
| # | Task | Labels |
|---|---|---|
| #17 | Slack app: interactive acknowledge and resolve, OAuth install | help wanted |
| #18 | Anomaly detection: seasonality-aware baselines and per-template thresholds | core team |
| #19 | RBAC: project-level roles and SSO group mapping | core team |
How work flows
- Pick an issue, comment to claim it, ask questions there.
- Branch from
main, keep PRs focused, add tests next to the code (see CONTRIBUTING.md). - CI must be green: typecheck, vitest, pytest, go test, mvn test, production
next build. - A maintainer reviews within a few days. Docs changes redeploy the site automatically.
Not planned
- A hosted free tier before 0.2 lands.
- Support for providers beyond OpenAI and Anthropic in the wrappers (use
regressa.trace()or OTLP instead) until demand shows up.