Tabby

Tabby (TabbyML) is a self-hosted AI coding assistant built by TabbyML, Inc., a San Francisco company founded in 2023 by Meng Zhang and Lucy Gao, both former Google AI researchers. The Rust-built server runs on-premise or in a private cloud with no external database or cloud service dependency, targeting engineering teams in banking, semiconductors, and defense where source code cannot leave the organization's infrastructure. TabbyML raised $7.2 million total and has accumulated over 32,000 GitHub stars since its April 2023 launch.

Key Capabilities:

• Self-contained server binary requiring no DBMS or cloud service, deployable via Docker, Kubernetes, or bare-metal

• RAG-based code completion that indexes local repositories for context-aware suggestions

• Answer Engine for team-level Q&A trained on internal codebases, documentation, and GitLab merge request history

• Support for consumer-grade GPUs running CodeLlama, StarCoder, CodeQwen, Codestral, and other open-source models through an official model registry

• IDE plugins for VS Code, JetBrains, Neovim, and Vim

• OpenAPI interface for integrating with existing CI/CD pipelines and cloud IDEs

• Admin UI with team management, LDAP authentication, usage analytics, and storage reporting

• GitHub and GitLab integration with branch, pull request, and CI/Lint/Test result visibility

• Shareable persistent pages from Answer Engine sessions

• Custom documentation ingestion through REST APIs

• Agent capabilities in private preview for organizations on the early access waitlist

See Tabby pricing details →