fix: ai chat

[Kitenite]

Description

Related Issues

Type of Change

• Bug fix
• New feature
• Documentation update
• Release
• Refactor
• Other (please describe):

Testing

Screenshots (if applicable)

Additional Notes

---

Important

Simplifies error handling in use-chat.tsx by removing redundant chat stopping logic in onError.

• Behavior:
  • Removes redundant chat stopping logic in onError function in use-chat.tsx.
  • Simplifies error handling by filtering out 'assistant' messages and updating conversation with last message.
• Misc:
  • Removes unnecessary timeout logic for stopping chat in use-chat.tsx.

^{This description was created by for 6043a1c. You can customize this summary. It will automatically update as commits are pushed.}

Summary by CodeRabbit

• New Features

  • Added support for the Claude 3.5 Haiku model with a 200k-token context, enabling longer prompts and responses.

• Bug Fixes

  • Improved chat error handling to avoid premature session termination, ensuring conversations aren’t cut off and state remains consistent.

[vercel]

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

Project	Deployment	Preview	Comments	Updated (UTC)
docs	Building	Preview	Comment	Sep 12, 2025 1:05am
web	Building	Preview	Comment	Sep 12, 2025 1:05am

[supabase]

This pull request has been ignored for the connected project wowaemfasoptxrdjhilu because there are no changes detected in apps/backend/supabase directory. You can change this behaviour in Project Integrations Settings ↗︎.

---

Preview Branches by Supabase.
Learn more about Supabase Branching ↗︎.

[coderabbitai]

Caution

Review failed

The pull request is closed.

Walkthrough

Removes immediate and delayed chat.stop() calls in error handling within the chat hook. Adds a new CLAUDE_3_5_HAIKU token limit entry (200000) to the LLM model token map. No exported/public signatures changed.

Changes

Cohort / File(s)	Summary
Chat error handling apps/web/client/src/app/project/[id]/_hooks/use-chat.tsx	Removed chat.stop() invocations in onError (both immediate and 100ms delayed). Retains handleChatError, assistant message filtering, and state updates.
LLM token limits packages/models/src/llm/index.ts	Added MODEL_MAX_TOKENS entry: [OPENROUTER_MODELS.CLAUDE_3_5_HAIKU]: 200000. No other changes.

Sequence Diagram(s)

sequenceDiagram
  autonumber
  actor User
  participant UI as Chat Hook (use-chat)
  participant Svc as Chat Service

  User->>UI: Send message
  UI->>Svc: chat.send(input)
  Svc-->>UI: Stream / Response

  alt Success
    UI->>UI: Update conversation state
  else Error
    UI->>UI: handleChatError(err)
    Note right of UI: Assistant messages filtered<br/>No chat.stop() called
  end

Loading

Estimated code review effort

🎯 2 (Simple) | ⏱️ ~10 minutes

Poem

A hiccup in chat, no sudden “stop!”
I twitch my ears, let errors pop.
New Haiku’s tokens, tall as hay—
Two hundred thousand hops today!
I thump, I test, I merge with cheer,
Bugs hop out, the path is clear.

---

📜 Recent review details

Configuration used: CodeRabbit UI

Review profile: CHILL

Plan: Pro

📥 Commits

Reviewing files that changed from the base of the PR and between 0908f7e and 6043a1c.

📒 Files selected for processing (2)

• apps/web/client/src/app/project/[id]/_hooks/use-chat.tsx (0 hunks)
• packages/models/src/llm/index.ts (1 hunks)

✨ Finishing touches

• 📝 Generate Docstrings

🧪 Generate unit tests

• Create PR with unit tests
• Post copyable unit tests in a comment
• Commit unit tests in branch fix/ai-1

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}