fauxpilot/fauxpilot

Health v2 · maintenance only

Flags: no_releases

Adoption not part of the score

14691 stars · 639 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-29, confidence not recorded

FauxPilot is a self-hosted, open-source alternative to the GitHub Copilot code completion server. It serves SalesForce CodeGen models via NVIDIA's Triton Inference Server with the FasterTransformer backend and exposes an OpenAI/Copilot-compatible API.

Use cases

• self-host a GitHub Copilot alternative for code completion
• serve code generation models locally on my own GPU
• point my editor's Copilot plugin at a private server
• run offline AI code suggestions without sending code to the cloud
• host an OpenAI-compatible completion API for code models

When to choose

• you want private, on-premise code completion with no code leaving your network
• you have NVIDIA GPUs with sufficient VRAM and Docker
• you want an OpenAI/Copilot-compatible API you control

When to avoid

• you lack an NVIDIA GPU or the VRAM for the models
• you want a polished, actively developed product with formal support
• you prefer managed cloud code assistants like GitHub Copilot

Facets

service · maturity maintenance

llm-inference api-framework http-server machine-learning developer-tools large-language-models artificial-intelligence self-hosted self-hosted python code-completion copilot-alternative codegen triton-inference-server fastertransformer local-ai docker linux gpu

Member repositories

For agents

markdown · JSON · MCP: product_card(name="fauxpilot/fauxpilot")