v2.0.32

Improve Autocomplete behavior to fix #1

v2.0.31

• Add new amazon/nova-premier-v1 model
• Remove missing free models: devstral-small, kimi-dev:72b, qwen3:8b
• Fix Z.ai provider not accepting metadata
• Update OpenRouter and Z.ai pricing
• Fix bg-opacity on small resolutions

v2.0.30

• Improved Responsive Layout with collapsible Sidebar. fixes #4
• Return focus to textarea after request completes. fixes #3
• Support VERBOSE=1 for enabling --verbose mode (useful in Docker)

v2.0.29

• Autoreload providers and UI config when change to config files is detected
• Add cancel button to cancel pending request
• Clicking outside model or system prompt selector will collapse it
• Clicking on selected item no longer deselects it

v2.0.28

• Configurable Request Size (limits/client_max_size 20MB)
• Auto Convert Large images (convert/image/[max_size 1536x1024,max_length 1.5MB])
• Dark Mode fixes

v2.0.27

• Dark Mode
• Drag n' Drop files
• Copy & Paste files

v2.0.26

Added support limiting access to GitHub OAuth Authenticated Users.

v2.0.25

• Improve support for Docker
• Auto generate ~/.llms/llms.json and ~/.llms/ui.json if not exist

v2.0.24

v2.0.24 Release Notes