Download Links (v1.11.2)
| OS | Architecture | Download URL |
|---|---|---|
| macOS | x64 | https://cdn.anythingllm.com/legacy/1.11.2/AnythingLLMDesktop.dmg |
| macOS | ARM64 (Silicon) | https://cdn.anythingllm.com/legacy/1.11.2/AnythingLLMDesktop-Silicon.dmg |
| Windows | x64 | https://cdn.anythingllm.com/legacy/1.11.2/AnythingLLMDesktop.exe |
| Windows | ARM64 | https://cdn.anythingllm.com/legacy/1.11.2/AnythingLLMDesktop-Arm64.exe |
| Linux | x64 | https://cdn.anythingllm.com/legacy/1.11.2/AnythingLLMDesktop.AppImage |
| Linux | ARM64 | https://cdn.anythingllm.com/legacy/1.11.2/AnythingLLMDesktop-Arm64.AppImage |
Key Features & UI Updates
- Workspace Config: Toggle workspace LLM provider and model directly from the main chat interface.
- Agent Controls: Toggle Agent skills on/off directly from the prompt input.
- Metrics & Citations: Added execution metrics and document/web-search citation reporting to Agent calls.
- Windows GPU Auto-Detection: Installer automatically detects GPU vendor (NVIDIA/AMD) on Windows x64 to download only hardware-matching GPU support binaries.
API & Integration Changes
- Ollama Embedder: Enabled long-timeout fetch configuration.
- Ollama API: Removed
use_mlockconfiguration parameter to resolve API compatibility issues. - Novita: Added native tool calling support.
- OpenRouter: Fixed stream metrics accuracy.
- Web Search Providers:
- Removed Google SERP (deprecated by Google).
- Added Perplexity Search SERP support.
Bug Fixes & Operational Notes
- LMStudio: Model picker now displays error states.
- Meeting Assistant: Model files are auto-verified prior to transcription to prevent execution-time failures.
- Chat Copy: Thinking process outputs are stripped from copied message content.
- Workspace Docs: Added document count indicators in the workspace document picker interface.
- Confluence: Applied wiki integration stability fixes.