Skip to content

v2.2.0

Latest
Latest

Choose a tag to compare

View all tags
@cuiyuebing cuiyuebing released this 03 Sep 03:25
v2.2.0
ae092fd
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

What's Changed

✨ Added

QwenPaw Hub

  • Self-host a multi-user QwenPaw Hub with local-process or Docker runtimes, workspace-level access controls, credential management, and reverse-proxy support (#7112)

QwenPaw Data

  • Deploy QwenPaw Data as a native PawApp with a durable analysis workspace, packaged runtime, and one-command Docker Compose example (#6940, #7190)

QwenPaw Mail

  • Add an intelligent mail assistant with real-time monitoring, search, thread-aware actions, sender allow/deny rules, and human approval for unknown recipients (#6800)
  • Bundle and launch the Mail MCP with the desktop app (#7170)
  • Keep Mail tool responses readable in Chinese conversations (#7303)

Models & Providers

  • Make model setup and selection more consistent with unified provider discovery, capability-aware routing, fallback policies, and thinking controls (#6302)
  • Add Volcengine Agent Plan and MiMo V2.5 providers, with refreshed Volcengine catalogs (#6515)
  • Add prompt-cache hit and write observability to Token Usage (#7342)

QwenPaw Creator

  • Creator 1.1.0 expands image and video generation across mainstream providers and improves narrative videos by carrying planned dialogue into generation prompts (#7167)
  • Creator 1.1.1 adds recorded website and desktop operation, Bailian Wan3 video generation, and configurable automated video-quality checks (#7274)

Console & Workspaces

  • Organize subagent conversations into groups and surface completion and unread status (#7035, #7275)
  • Make background work easier to follow with a compact task list and consistent run tracking (#7083, #7319)
  • Show response artifacts and make generated media directly downloadable (#7161, #7036)
  • Select system-prompt files from the workspace UI (#7078)
  • Create folders and copy files from the workspace UI (#7151, #7314)
  • Keep project directories scoped to each session so concurrent work can use different roots safely (#6976)
  • Expand Token Usage with model and tool-call trends, per-Agent breakdowns, and an all-Agent view (#7207, #7219)
  • Auto-hide the QwenPaw OS dock for maximized windows (#7233)
  • Improve navigation when editor tabs overflow (#6983)
  • Copy assistant responses without including hidden reasoning text (#7448)

Skills, Channels & Media

  • Integrate AnySearch as a web SearchProvider and MCP service (#7081)
  • Search Skills and enable or disable them in batches (#7140, #6990)
  • Keep Skills current with faster file loading, cleaner bindings, and automatic updates for built-ins (#7097, #7232)
  • Manage apps, plugins, and Skills in one marketplace (#6880, #7017, #7235)
  • Open newly installed PawApps without reloading the Console (#7238)
  • Configure plugin channels interactively from the CLI (#6943)
  • Include custom channels in Agent MCP access rules (#7225)
  • Keep Matrix group sessions and memories isolated per sender, with retries for transient login failures (#7001, #7109)
  • Let DingTalk groups share conversation context when configured (#7208)
  • Configure automatic layout for DingTalk widescreen cards (#7416)
  • Download OneBot media locally before Agent processing (#6715)
  • Let Computer-use observe related windows (#7037)
  • Optionally resize oversized images before sending them to a model (#7294)
  • Propagate Console stop requests to active Agent tools (#7349)
  • Automatically fold completed assistant process messages (#7374)

🔄 Changed

  • Start persistent Drivers concurrently for faster workspace startup (#7174)
  • Reduce streaming overhead during long responses (#7244, #7239)
  • Stabilize reusable prompt-cache prefixes across requests (#7346)
  • Keep long, Markdown-heavy chats responsive (#7176)
  • Use tighter, more consistent spacing for Markdown lists (#7317)
  • Improve the ReMe runtime dashboard and long-term-memory guidance (#6984, #6997)
  • Stop sending notifications when a scheduled job has not changed (#7115)
  • Remove obsolete Desktop Mode reminders and outdated approval hints (#7245, #6981)
  • Update AgentScope through 2.0.7.post1 (#7276, #7373)
  • Refresh the Console chat package used by this release (#7257, #7289)
  • Clarify the behavior of each approval mode (#7368)
  • Make embedding-model changes explicit and reversible, falling back to BM25 until a scoped reindex succeeds (#7133)
  • Rename Marketplace to Extensions and restore dark-mode styling across the Console (#7454)

🐛 Fixed

Models, Runtime & Memory

  • Restore complete free-model listings and legacy model names (#7175)
  • Refresh Aliyun and Kimi model catalogs (#7271, #7277)
  • Separate model output capabilities from request limits so routing and validation use the correct metadata (#7337)
  • Improve compatibility with strict DashScope tool schemas and audio messages (#7284, #7024)
  • Honor server-provided retry limits during streaming (#6617)
  • Recover from stalled model streams (#7150)
  • Propagate and report model timeouts accurately (#7268, #7308)
  • Keep Agent behavior compatible with AgentScope 2.0.6 (#6908)
  • Enable the Ollama embedding backend (#7131)
  • Preserve embedding verification while navigating (#7227)
  • Exclude reasoning text from generated chat titles (#7187)
  • Give background tasks a reliable default timeout without overriding explicit values (#6869)
  • Start ReMe memory services before model configuration so local memory features remain available without a configured model (#7468)
  • Include ~-based Agent workspace paths in model and tool-usage trends (#7432)

Console, Files & Media

  • Reopen the sole history session directly from New Chat (#7104)
  • Show the actual completion time for assistant responses (#6938)
  • Render historical data-URL images after reload (#7069)
  • Improve model-selector styling and dark-mode session groups (#7137, #7111)
  • Hide empty grep badges and unavailable approval controls (#7165)
  • Improve mobile composer controls (#7334)
  • Align cron calendar dates with Sunday-first headers (#7352)
  • Keep embedding verification when switching Agents (#7353)
  • Preserve uploaded and non-ASCII filenames (#6492, #7192)
  • Preserve Windows CRLF line endings and refresh file previews after changes (#7091, #7094)
  • Continue remote-file download fallback after a downloader timeout (#6371)
  • Persist remote images before saving tool results (#7146)
  • Deliver tool-result videos through the OpenAI Responses API (#7061)
  • Make inline video limits configurable (#7071)
  • Preserve local file URLs while normalizing media messages (#7343)
  • Save screenshots to the active project directory (#7439)
  • Link Daily Paper to the correct QwenPaw documentation (#7466)

Channels

  • Restore Yuanbao message decoding with protobuf 6 and later (#7421)

Plugins, Desktop & Operations

  • Restore plugin workspace state before reloads and isolate absolute imports between plugin namespaces (#6996, #6688)
  • Recover Windows WebView2 after a browser crash (#7128)
  • Avoid false process matches in the Windows installer, including the NSIS caller itself (#7281, #7323)
  • Handle NSIS uninstall blockers and make application-data cleanup choices explicit (#7336, #7354)
  • Clean up safely when workspace startup is cancelled (#7194)
  • Protect environment configuration from partial or corrupt writes (#7135)
  • Keep backup jobs running after an SSE connection closes (#7283)
  • Keep Agent configuration persistence consistent across shared filesystems and concurrent writes (#6767)
  • Enforce File Guard paths during active policy evaluation (#7375)
  • Expand sandbox mount paths correctly (#7116)
  • Avoid antivirus interference during sandbox setup (#6986)
  • Isolate QQ conversations and preserve their reply routes (#7169)
  • Report Xiaoyi proactive-send failures reliably and preserve session IDs (#7147)
  • Patch security-sensitive website and Creator dependencies (#7172)
  • Speed up plugin loading by avoiding a full module scan after every plugin (#7383)
  • Restore ReMe and its Auto Fin and Daily Paper plugins in packaged desktop builds (#7458)
  • Prevent desktop packages from mixing different Python runtime sources (#7372)
  • Block shell line-continuation variants from bypassing sensitive-path protections (#7472)

New Contributors

Full Changelog: v2.1.0...v2.2.0

Assets 4
Loading