Sidecar is the AI backend for the Aide code editor, providing local, on-machine intelligence for prompt generation and LLM interaction. It's designed for developers seeking to integrate advanced AI capabilities directly into their coding workflow without relying on external cloud services.

How It Works

Sidecar leverages a modular architecture with distinct components: tool_box.rs for AI tool management, a symbol/ directory for granular code understanding (from files to functions), and repomap which generates a repository map using a PageRank-like algorithm for code symbol analysis. This approach allows for flexible and context-aware AI assistance tailored to individual codebases.

Quick Start & Requirements

• Install/Run: Build the binary with cargo build --bin webserver and run it with ./target/debug/webserver.
• Prerequisites: Rust 1.79.
• Setup: Building from source is required. Integration with Aide editor involves running the Sidecar binary alongside Aide.

Highlighted Details

• Acts as the AI "brains" for the Aide editor.
• Manages AI tools and handles LLM interactions locally.
• Features a versatile symbol analysis system for code understanding at various granularities.
• Includes a repository mapping component inspired by Aider's Python implementation.

Maintenance & Community

• Project welcomes contributions via bug reports, feature requests, and code reviews.
• Community engagement is encouraged through Discord.
• Adheres to the Contributor Covenant Code of Conduct.

Licensing & Compatibility

• Licensed under the GNU Affero General Public License v3.0.
• The AGPLv3 license may impose copyleft restrictions on derivative works, potentially impacting commercial or closed-source integrations.

Limitations & Caveats

The inline completion features are noted as not actively maintained, though contributions are welcomed. The AGPLv3 license requires careful consideration for commercial use cases.