What Fauxpilot Local does
FauxPilot is described on its repository as an open-source alternative to a GitHub Copilot server, built to run locally. It uses Salesforce CodeGen models inside the NVIDIA Triton Inference Server with the FasterTransformer backend. Clients can use the OpenAI API, a Copilot plugin or a REST API. Setup uses a script that lets the user choose a model, downloads it in GPT-J format and converts it for FasterTransformer. Prerequisites are Docker, docker compose 1.28 or later, an NVIDIA GPU with Compute Capability 6.0 or later, nvidia-docker, curl and zstd. Separate guides cover server and client setup. The license is MIT, and the page does not state pricing.
How Fauxpilot Local compares with its closest alternatives
4 of the 5 closest alternatives have a free plan, as Fauxpilot Local does. Among the 2 that list a starting price, the cheapest is Claude Code at $20 a month and the median is $20. 1 of the 5 offers an API, as Fauxpilot Local does. Fauxpilot Local was founded in 2023; the alternatives were founded between 2021 and 2024, 2 of them earlier. Fauxpilot Local has no single headquarters on record; its alternatives with a known headquarters are in the United States.