Spark

SparkLang roadmap

Prioritized against the adoption bar. Status: done / next / won't (near-term product focus). Details: ADOPTION_BAR.md.

Done (this era)

  • Public MIT repo, SparkLang branding, About disambiguation
  • Changelog + release process doc
  • embed / retrieve first-class ops + dry fixtures + spark-rag-http
  • Dry-run escape hatch: shell / run allowlist fixtures
  • ./spark --embed FFI handshake stub
  • Per-ask [accounting] line (dry zeros; live usage when present)
  • Positioning: not a Bifrost plugin; dry-run first
  • http get / http post + timeout + dry fixture files + live ./spark-http (auth/retries still next)

Next

Track What
HTTP Auth / bearer headers; retries
Extract Schema validation + JSON-mode + retry on miss
Ask Streaming token/SSE path
Eval Expectation pass/fail harness (not only alias compare)
Escape Live --allow-shell argv policy; import spark host embed
Accounting Wall-clock latency on live ask; run-level rollup
LSP Prefer editor LSP + highlighting over IDE chrome
Receptionist Promote examples/receptionist_goal.spark off [goal] when transfer + eval syntax exist
Releases Cut GitHub Release tags from CHANGELOG (see RELEASE.md)

Done (rename)

Track What
Rename Public sparklang-dev/sparklang live; private Michael20251/sparklang retained. Site sparklang.dev. CLI remains spark.

Won't (near-term focus)

Drop Why
MITM / pcap / browser automation as hero Distracts from AI workflow language
Homegrown Electron/PyQt IDE as hero chrome Prefer LSP for editor story; IDE tree kept (hard-delete OWNER-CONFIRM revoked 2026-09-02)
Model-build as training product Blueprint / eval sugar only
Production voice/telephony claim Gaps remain (barge-in, EOU, SM, adapters) — gated demo only
Bifrost-as-identity Gateway optional; any OpenAI-compatible URL

Voice gaps (honest)

If not shipping soon, keep voice off the production claim list:

  • Barge-in, EOU / silence detection, latency budget
  • Call state machine: transfer / hold / hangup / VM / DTMF / STT retry
  • Named STT / TTS / telephony adapters as stable contracts
  • Concurrency model for multi-call

Dry listen / speak / gated PSTN remain demos. See VOICE.md.