Skip to content

Utterform 0.10.0 Text actions on your ChatGPT subscription - #50

Merged
jli-software merged 2 commits into
mainfrom
work/utterform-0.10.0-codex
Nov 7, 2026
Merged

jli-software merged 2 commits into
mainfrom
work/utterform-0.10.0-codex

Conversation

@jli-software

Copy link
Copy Markdown
Owner

Stage 0.10.0 of docs/plans/v1.0-roadmap.md. Ships with 1.0; no tag, no release.

  • Settings → AI → Provider: OpenAI key | ChatGPT (Experimental). Transcription unchanged.
  • transcription/codex.rs: Codex CLI login, token renewal with atomic write-back, SSE answer, model list from models_cache.json, Auto = fastest.
  • No fallback to the API key (decided 2026-11-07); a missing login stops a recording before it starts.
  • Live check on asus16: GPT-6-Luna 1.5 s vs gpt-5-mini over the API 5–7 s.

Checks: npm run check, npm test (145), npm run build, Playwright (35), cargo fmt --check, cargo clippy -D warnings, cargo test (189).

🤖 Generated with Claude Code

https://claude.ai/code/session_014LfoKbq3S36AiYRW7j8QLK

jli-software and others added 2 commits November 7, 2026 14:15
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_014LfoKbq3S36AiYRW7j8QLK
Settings → AI → Provider chooses OpenAI key or ChatGPT. With ChatGPT the
text actions run through the Codex CLI login (~/.codex/auth.json) on
chatgpt.com/backend-api/codex, renewing an expiring token and writing it
back for the CLI. Transcription is unchanged. Marked experimental; no
fallback to the API key. Models come from the CLI's model cache, Auto
takes the fastest.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_014LfoKbq3S36AiYRW7j8QLK
@jli-software
jli-software merged commit d06688a into main Nov 7, 2026
2 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant