Roadmap
Roadmap
Roadmap
Quarter-level view. Live task status lives in the
chttp project; this page records
milestones and scheduling, not per-task state.
2026-Q4
M1: Ecosystem expansion (October)
Grow the package family around @chttp/router with validation and build-time
adjuncts. Everything here reuses existing seams: the val schema surface and
the unplugin route extraction. The route contract stays frozen.
| Workstream | Effort | Focus |
|---|---|---|
JSON Schema emission for @chttp/val |
M | every public schema and pipe action maps to JSON Schema 2020-12; valid OpenAPI 3.1 by construction |
OpenAPI side-channel emit in @chttp/unplugin |
M | opt-in .chttp/routes.openapi.ts assembled from the extracted route table plus statically locatable val schemas |
| Typed client emit | S/M | .chttp/routes.client.ts from the same extraction; types from the build artifact, not chained generics |
| Env validation example | S | recipe on top of @chttp/val, not a package |
| Agent skills, middleware cookbook, deploy template | S | docs and copyable templates |
M2: Credibility (October to November)
- Deploy story promotion: scratch-container images, cold-start and RSS claims
backed by reproducible artifacts. - Bench matrix upkeep. The compiled-Hono comparison cell is documented as
blocked: scriptc 0.3.1 cannot serve Hono (its core dispatch is a fenced
site); evidence transcripts inbench/scriptc-evidence/. Revisit when
scriptc releases widen coverage.
M3: Frontend hypothesis (Q4)
Explored in a parallel track: an html tagged-template recipe, islands on top
of JSON endpoints, SSG prerender as an unplugin pass. This is a demand probe.
The gate for any framework-scale commitment is evidence from the probe, not
ambition. No M1/M2 package depends on it.
Last updated:
Recent commits
