Skip to content

Version Packages - #1248

Merged
mikebarkmin merged 1 commit into
mainfrom
changeset-release/main
Nov 16, 2026
Merged

mikebarkmin merged 1 commit into
mainfrom
changeset-release/main

Conversation

@github-actions

@github-actions github-actions Bot commented Nov 15, 2026 •

Copy link
Copy Markdown
Contributor

This PR was opened by the Changesets release GitHub action. When you're ready to do a release, you can merge this and the packages will be published to npm automatically. If you're not ready to do a release yet, that's fine, whenever you add more changesets to main, this PR will be updated.

Releases

@hyperbook/cloud@0.12.0

Minor Changes

  • 7032a19 Thanks @mikebarkmin! - The editor's AI assistant works like a full chat now. Conversations are kept per teacher and book: they survive a reload, are listed in the pane's history, can be renamed and deleted, and a follow-up sees the whole conversation, not just the last few answers cut to 4000 characters. A reasoning model's thinking streams live into a "Thinking…" block that folds away once the answer starts, and a tool call shows its progress while it is being written ("Writes arbeitsblatt-2.typ — 6.2 KB"), so a long answer no longer looks stuck. An answer cut off by the output cap goes on by itself: cut-off text is continued, a cut-off file write is retried in smaller pieces, and thinking that used up the cap is told to act — thinking is never passed off as the answer any more. When the step limit ends unfinished work, Continue picks it up. File changes the teacher has not accepted or denied yet carry over, so "make task 5 harder" changes the pending draft. Messages can be edited and resent, answers regenerated or retried, and images, PDFs and text files attached (button, paste or drop); a picker sets how hard a reasoning model thinks. The provider timeout is an idle timeout now — it fires only when the provider goes silent, never while it streams — and the defaults grow to 65536 output tokens (only what is written is billed; a provider that rejects that is asked again with 16k) and 12 steps. Needs a database migration (applied on start).

Patch Changes

  • 9905d48 Thanks @mikebarkmin! - The AI assistant shows an answer in the order it happened: each stretch of the model's thinking gets its own folding block between the files it read and the text it wrote, instead of one thinking block at the top that kept filling up while the model worked. The order is kept with the conversation, so it looks the same after a reload. The history and new-chat buttons show their pointer and tooltip again (they are no longer disabled — switching chats stops a running answer), and the thinking-effort picker no longer touches the input field. Needs a database migration (applied on start).

@vercel

vercel Bot commented Nov 15, 2026 •

Copy link
Copy Markdown

The latest updates on your projects. Learn more about Vercel for GitHub.

Project Deployment Actions Updated
hyperbook-4fwa Ready Ready Preview Nov 15, 2026 10:19pm UTC

@github-actions
github-actions Bot force-pushed the changeset-release/main branch from 759db10 to 56de9fb Compare November 15, 2026 22:15
@mikebarkmin
mikebarkmin merged commit 4496848 into main Nov 16, 2026
2 checks passed
@mikebarkmin
mikebarkmin deleted the changeset-release/main branch November 16, 2026 05:38

This branch was successfully deployed

1 active deployment
Preview — 56de9fb4 Deployed Nov 15, 2026 by vercel[bot]
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