Changelog:

• Fixed crash on Android 14+
• New models: SmolLM3, SmolVLM2, Qwen3.5, LFM2.5, LFM2.5-VL, Ultravox
• Multimodal support: Vision input via camera or gallery, Audio input for transcription and Q&A
• Characters: create, save and switch between custom AI personas via system prompts
• Chat management: save and delete conversations
• Syntax highlighting in code blocks with copy button
• Expert Mode: import external models (gguf), expanded KV cache up to 32768 tokens, max tokens up to 16384, model info viewer
• Bug fixes and stability improvements

Changelog:

• Major UI/UX enhancements (includes fixes for flickering and persistent notifications)
• Added support for model variants and new models (Qwen3, Google Gemma3)
• Introduced uncensored models (use with caution)
• Expert Mode: adjustable temperature, top-k/top-p, kv-cache and token limits
• Model management tools: download progress, deletion
• Benchmark testing added to assess model performance metrics

Changelog:

• Fixed an issue where downloaded models were not displaying
• Minor UI improvements

AbodeLLM 1.2.1 Changelog

• Added a smaller model (SmolLM2) for better inferencing on low-end devices
• Minor bug fixes

AbodeLLM 1.2.0 Changelog

• Added new model (DeepSeek-R1-Qwen)
• Further Increased Token Size to allow for longer responses and more context
• Added Response Formatting (bold, italics etc.)
• Improved Responses for LLaMA model
• Added Auto Context Keeping to balance context remembered and model's token limit
• Reduced flickering when generating response
• UI Improvements

AbodeLLM 1.1.3b Changelog

• Token limit notifier
• Minor bug fixes

AbodeLLM 1.1.2b Changelog

• Increased Token Size to allow for longer responses and more context.
• Improved Brief Responses
• Improved Context Keeping
• Minor UI Fixes

AbodeLLM 1.0.2b

Key Features:

• Local AI Processing: The app operates entirely on your phone, ensuring that all data remains private.

• Simple and Unlimited Responses: Users can receive answers without daily limits.

Known Issues and Limitations:

• Context History: The feature is not yet available, meaning the chatbot may truncate responses or fail to recall earlier parts of the conversation.
• Performance: While the app is functional on most devices, it is recommended to use a device with higher specifications for improved response times.

AbodeLLM 1.1.0b Changelog

New Features:

• Updated user interface for an enhanced user experience.

• Added context feature (can be enabled in settings) to maintain conversation continuity.

• Introduced short responses option (disabled by default) for quicker replies.

• Added option to clear chat for a fresh start.

Improvements:

• Enhanced memory management for better performance and efficiency.

Bug Fixes:

• Implemented minor bug fixes to improve overall app stability.

AbodeLLM 1.0.4b Changelog

New Features:

• Added a notification to prompt users to stop the app when it is running in the background to conserve memory.

Bug Fixes:

• Fixed crashing issues specifically on Android Nougat.

• Resolved memory leaks to improve overall stability.

• Addressed freezing issues that occurred after downloading the model.