desktop-v0.0.12
ClineSummary
This release overhauls the transcript experience with individual tool call rows, terminal-like command display, and inline diffs for edits. Key bug fixes include improved daemon handling and corrected behavior for strict providers, while enhancements focus on typography, UI layout, and more accurate token usage for model reasoning.
New Features
- Introduced individual rows for each tool call in the transcript, featuring distinct icons, statuses, and expandable details for better clarity.
- Enabled terminal-like command display with captured output upon expansion for a more intuitive experience.
- Added inline diff views for edits, showing changes hunk by hunk.
- Enabled the Claude Code provider for agentic work, anchoring sessions to the workspace folder, loading settings, and allowing file edits within the workspace.
- Added Fable 5 and custom model IDs to Vertex AI, and improved the global-region picker to display all models from the live catalog.
Bug Fixes
- Fixed an issue where reconnecting to a stale managed Cline Hub daemon would cause problems; the app now prompts to update and restart correctly.
- Resolved an issue where the Hub daemon would shut down uncleanly when a client was still connected.
- Fixed strict providers rejecting turns with an "user message must have content" error when the message contained only empty text.
- Fixed a mid-turn crash that occurred with streamed tool calls having non-zero indexes, particularly when using LiteLLM's Anthropic passthrough.
- Fixed the first turn of a fresh session from getting stuck on "Agent is working…" indefinitely.
- Resolved an issue where truncated tool-call JSON was being silently repaired instead of being rejected.
Improvements
- Enhanced the session transcript layout, message surfaces, and composer with new Inter and Geist Mono typography.
- Updated the thinking indicator to remain active during quiet stretches of a turn, preventing the turn from appearing frozen.
- Redesigned message actions (copy, fork, timestamp) to no longer overlap with message text, and expanded panels now render at full opacity.
- Centered and top-aligned the chat input on the welcome screen and temporarily hid prompt suggestions.
- Ensured scheduled runs no longer appear in the session sidebar and history list.
- Updated file diff rendering to use a shared, syntax-highlighted renderer that respects the app's theme.
- Modified compaction to respect the Max Output Tokens setting instead of a hardcoded 1024-token cap, allowing reasoning models to utilize their full budget.