Sidekick LLM

Features are:

• Use images and files in your prompt (drag and drop or copy paste)
• Local history of all conversations with all models
• Continue any conversation from any point from the history tab!
• Local LLM Arena! Blind model comparison with ELO rating
• "Thinking mode", inspired by o1. Let the model chain together an internal monologue before answering.
• Didn't like the response? Regenerate with a different model!
• Pop out Sidepanel into a new tab and vice versa
• Customizable system prompts that will give you better and more fun responses
• Control model temperature and other settings
• Incognito mode (can be toggled mid conversation)
• Automatically title your chats using any LLM
• Add new models in the settings tab
• Token tracking
• Web search tool for supported models
• NEW:
• Council Mode: Let multiple LLMs respond at once, then one "Chief" LLM refines it into a final response
• Tabs
• Now also supports local models (llama cpp and vllm)
• Added seamless support for nano banana
• Search your history
• Dedicated tab for images in history

Currently supports OpenAI, Anthropic, Gemini, DeepSeek, Grok, etc. models. Requires personal API keys for model access.

For more detailed information on how to use this extension, if you have any issues, or want to contribute, please visit: github.com/EduardR02/Faster-ChatGPT