travisennis / acai

An AI coding assistant.

acai

Acai is a versatile AI-powered coding assistant that integrates with your development workflow. It provides intelligent code suggestions, documentation generation, and interactive chat capabilities to enhance your coding experience.

Features

• Interactive chat mode for coding assistance
• Code completion and suggestions
• Code optimization suggestions
• Automatic documentation generation
• Bug fixing and code improvement suggestions
• Support for multiple AI providers (OpenAI, Anthropic, Mistral, Google, Ollama)
• Language Server Protocol (LSP) integration
• Customizable prompts and templates
• File filtering for codebase analysis

Installation

To install Acai, you'll need Rust and Cargo installed on your system. Then, follow these steps:

1. Clone the repository:

   git clone https://github.com/travisennis/acai.git
   cd acai

2. Build the project:

   cargo build --release

3. The binary will be available in target/release/acai

Usage

Acai provides several subcommands for different functionalities:

• chat: Start an interactive chat session with the AI assistant
• instruct: Send a single instruction to the AI
• lsp: Start the Language Server Protocol server

Chat Mode

Start an interactive chat session with the AI:

acai chat --model anthropic/claude-3-sonnet-20240229 --path /path/to/your/codebase

Instruct Mode

Generate code or documentation based on instructions:

acai instruct --model openai/gpt-4 --prompt "Implement a binary search tree in Rust" --path /path/to/your/codebase

Prompt Generator

Generate prompts based on your codebase:

acai prompt-generator --path /path/to/your/codebase --include "**/*.rs" --exclude "target/**"

LSP Mode

Start Acai as a Language Server:

acai lsp

Configuration

Acai uses environment variables for API keys:

• OPENAI_API_KEY: For OpenAI models
• CLAUDE_API_KEY: For Anthropic models
• MISTRAL_API_KEY: For Mistral models
• GOOGLE_API_KEY: For Google models

You can also customize the behavior using command-line options. Refer to the help command for each subcommand for more details:

acai <subcommand> --help

Example

acai chat --model openai/gpt-4 --temperature 0.7

Contributing

Contributions to Acai are welcome! Please follow these steps:

1. Fork the repository
2. Create a new branch for your feature or bug fix
3. Make your changes and commit them with clear, descriptive messages
4. Write tests for your new functionality
5. Submit a pull request

Make sure to run tests and linters before submitting:

cargo test
cargo clippy

Testing

To run the test suite:

cargo test

License

Acai is licensed under the MIT License. See the LICENSE file for details.

Acknowledgements

Acai uses several open-source libraries and AI models. We're grateful to the developers and organizations behind these technologies:

• Rust and the Rust community for providing excellent tools and libraries that make projects like this possible.
• OpenAI, Anthropic, Mistral, and Google for their AI models.
• The developers of crates used in this project (tokio, clap, reqwest, and others). Please see the Cargo.toml file for a full list of dependencies.