GitHub

🐾 Tabby

What is the project about?

Tabby is a self-hosted AI coding assistant, designed as an open-source and on-premises alternative to GitHub Copilot.

What problem does it solve?

It provides AI-powered code completion and assistance without relying on external cloud services or databases, giving users more control over their data and infrastructure. It offers a solution for developers and teams who prefer or require a self-hosted solution for code completion.

What are the features of the project?

  • Self-contained: No need for a DBMS or cloud service.
  • OpenAPI interface: Easy integration with existing infrastructure (e.g., Cloud IDE).
  • Consumer-grade GPU support.
  • Easy to deploy with Docker.
  • IDE/Editor Extensions.
  • RAG-based and locally relevant code completion.
  • Admin UI and team management.
  • Answer Engine for internal engineering teams.
  • Chat in side-panel and editing via chat command.
  • Seamless integrations (Gitlab SSO, Self-hosted GitHub/GitLab, etc.).
  • Flexible configurations (HTTP API integration).
  • Auto-generated commit messages.
  • Notification Box.
  • LDAP Authentication.

What are the technologies used in the project?

  • Docker (for deployment).
  • Rust (build environment).
  • Protobuf, OpenBLAS.
  • Various AI models (e.g., StarCoder, CodeLlama, Qwen2).

What are the benefits of the project?

  • Privacy and Control: Self-hosted nature ensures data remains within the user's infrastructure.
  • Open Source: Transparency and community contributions.
  • Cost-Effective: Potentially lower costs compared to cloud-based solutions, especially with consumer-grade GPU support.
  • Integration: OpenAPI interface allows for easy integration with existing tools.
  • Customization: Ability to choose and configure AI models.

What are the use cases of the project?

  • Individual developers seeking a self-hosted coding assistant.
  • Teams and organizations requiring on-premises code completion for security or compliance reasons.
  • Companies with existing infrastructure that want to integrate AI code completion without relying on external cloud services.
  • Internal engineering teams using Answer Engine.
tabby screenshot