[mrubens]

Promote v1.19.4

Frozen at eb9dc4b46d9deae3df13946eb4f873f86f4056af — the commit where 1.19.4 was versioned. Commits merged to develop afterward require an explicit candidate refresh or ship in the next release.

• Merge this PR with a merge commit (do not squash or rebase).
• If multiple promote PRs are open, merge them in version order (oldest first).
• Monitor CI and reviews on this PR's current head, not only the release-preparation PR. Unresolved findings and missing required approvals remain blockers even with green CI.
• After a refresh or CI reconciliation, re-check the new head; monitoring never authorizes merging or publication.
• Merging publishes a GitHub Release for v1.19.4 and triggers the existing GHCR v* image publish (latest channel).
• The release/v1.19.4 branch can be deleted after this PR merges.

Changelog

1.19.4 (2026-10-27)

Roomote 1.19.4 expands model choice with GPT-6.1 Terra and Kimi K3.1 Pro, including Highspeed options on supported Kimi providers.

Highlights

• Use GPT-6.1 Terra across supported API and subscription providers, with Fast mode on eligible ChatGPT plans.
• Choose Kimi K3.1 Pro or its Highspeed variant through supported Kimi providers and gateways.

Patch changes

• Select GPT-6.1 Terra through Roomote inference, OpenRouter, Vercel AI Gateway, OpenAI, Azure OpenAI, Azure AI Foundry, Requesty, and ChatGPT Subscription, including Fast mode on eligible ChatGPT plans. Existing defaults and saved model mappings stay unchanged.
• Select Kimi K3.1 Pro through Moonshot AI, OpenRouter, Vercel AI Gateway, Kimi for Coding, and OpenCode Go, with Highspeed available through Moonshot AI, Vercel AI Gateway, and eligible Kimi for Coding memberships.

[roomote-community]

No code issues found. See session

_{Reviewed eb9dc4b}