About Shore
Shore is a terminal based chatbot interface that is simultaneously minimalist, productive, and aesthetically pleasing.
Core Features
Section titled “Core Features”- Terminal based with vim inspired key bindings - no need to leave your terminal for LLM chats
- Chat concurrently with multiple models and providers - Create a new chat with one specific model, or with 20 different models spanning multiple LLM providers
- Free and open source - specific license TBD
- Support for most major LLM providers (see below) - eligible providers automatically detected based on API keys present in shell environment variables
- Locally stored chat history and settings are all stored locally
- Full text search - quickly review learnings from prior conversations before submitting new prompts
- Blazingly Fast - written 100% in Rust 🦀
Terminal based with vim inspired bindings
Section titled “Terminal based with vim inspired bindings”Use n to create a new chat and put the prompt field into insert mode. The prompt entry field behaves like a vim buffer. Press Esc to return to normal mode. Use b and w to traverse words. cc deletes the entire line and puts you back in insert mode. Use h and l to switch between models associated with the current chat. j and k traverse down and up in the current chat messages, respectively. These are just a few examples - see key bindings for a comprehensive list.
Concurrent Chat
Section titled “Concurrent Chat”Use one model for a traditional chat experience. Add additional models to a chat to send prompts all models concurrently. Use h and l to cycle through the responses of each model.
Free and open source
Section titled “Free and open source”Contributions are welcome! Shore GitHub Repository
Supported LLM Providers
Section titled “Supported LLM Providers”- Hugging Face
- OpenAI
- Anthropic
- Groq
- xAI
- Perplexity
- MiniMax
- zAI
Locally Stored
Section titled “Locally Stored”All conversation history, chat profiles etc are stored in a local SQLite database. default location is ~/.shore/default.db
Full Text Search
Section titled “Full Text Search”In normal mode, use / to enter search mode. Search terms will be matched on both titles and content of your chat history.