# ravenscroftj/turbopilot

Turbopilot is an open source large-language-model based code completion engine that runs locally on CPU

Repository: https://github.com/ravenscroftj/turbopilot
Canonical: https://ross.abutalabs.com/products/turbopilot
Language: C++
License: BSD-3-Clause
License Family: permissive
Topics: code-completion, cpp, language-model, machine-learning
Archived: true
Last push: 2023-09-30T08:16:59+00:00

## Health v2 (maintenance only)
Score: 10/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 0, release rhythm 8, longevity 88
- inputs: {"age_days": 1242, "days_push": 1068, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: archived
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 3779, forks 121 (observed 2026-08-28T04:08:18.370365+00:00)

## What it is
TurboPilot is a self-hosted, Copilot-like code completion server that runs large language models (Salesforce CodeGen, StableCode, WizardCoder, Starcoder) locally on CPU using llama.cpp's GGML library. It exposes an API compatible with Copilot plugins and runs in as little as 4GiB of RAM.

## Use cases
- self-hosted code completion without sending code to the cloud
- run a copilot clone on a CPU-only machine with limited RAM
- serve local codegen models to editor plugins
- experiment with quantized code LLMs like StableCode and WizardCoder
- replace GitHub Copilot with a private offline alternative

## When to choose
- you need fully offline, private code completion on modest hardware
- you want to experiment with local code LLM inference in C++

## When to avoid
- you want a maintained tool - the project was archived in September 2023
- you need fast, production-quality autocomplete
- you have a GPU and can use more mature alternatives

## Facets
- artifact type: application
- maturity: abandoned
- function: llm-inference, http-server, developer-tools
- domain: large-language-models, developer-tools, self-hosted
- platform: cpp, self-hosted, windows
- tags: code-completion, copilot-clone, llama-cpp, codegen, cpu-inference, deprecated, docker, linux, macos

## Member repositories
- ravenscroftj/turbopilot (main) score 10

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:08:18.370365+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T18:27:47.178295+00:00, confidence not recorded.
  - readme: https://github.com/ravenscroftj/turbopilot (fetched 2026-08-28T04:08:18.370365+00:00, sha a23481d064d3)
- Data as of 2026-08-30T08:39:29.467469+00:00.