# OpenPipe/OpenPipe

Turn expensive prompts into cheap fine-tuned models

Repository: https://github.com/OpenPipe/OpenPipe
Canonical: https://ross.abutalabs.com/products/openpipe
Homepage: https://openpipe.ai
Language: TypeScript
License: Apache-2.0
License Family: permissive
Topics: ai, llm, llmops, prompt-engineering
Last push: 2024-05-25T00:18:13+00:00

## Health v2 (maintenance only)
Score: 29/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 0, release rhythm 35, longevity 83
- inputs: {"age_days": 1171, "days_push": 831, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 2830, forks 178 (observed 2026-08-28T04:07:24.308616+00:00)

## What it is
OpenPipe is an open-source fine-tuning and model-hosting platform that turns expensive LLM prompts into cheaper fine-tuned models. It offers OpenAI-compatible SDKs and endpoints, request logging, dataset management, and model evaluation.

## Use cases
- reduce llm api costs by fine-tuning a smaller model
- fine-tune mistral or llama on my own prompt data
- log and query openai chat completion requests
- compare fine-tuned model output against gpt-3.5-turbo
- swap gpt-4 for a fine-tuned model with one line of code
- import openai jsonl datasets for training

## When to choose
- you want to cut inference costs by distilling a large model into a fine-tuned smaller one
- you need OpenAI-compatible APIs so switching models requires only a model-name change
- you want request logging, dataset pruning, and model evaluation in one platform

## When to avoid
- you need actively developed open-source code - development is paused pending an open-core restructure
- you only need prompt engineering or orchestration without fine-tuning
- you require fully self-hosted training on your own infrastructure without platform dependencies

## Facets
- artifact type: service
- maturity: maintenance
- function: llm-training, llm-inference, prompt-engineering, sdk
- domain: large-language-models, machine-learning, developer-tools, artificial-intelligence
- platform: self-hosted, python
- tags: llmops, fine-tuning, openai-compatible, model-hosting, cost-optimization, web-server, nodejs, docker

## Member repositories
- OpenPipe/OpenPipe (main) score 29

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:07:24.308616+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 2026-08-30T07:38:10.179154+00:00, confidence not recorded.
  - readme: https://github.com/OpenPipe/OpenPipe (fetched 2026-08-28T04:07:24.308616+00:00, sha 33039ad3398b)
  - homepage: https://openpipe.ai (fetched 2026-08-29T09:53:34.557786+00:00, sha 8ac58e9a2301)
- Data as of 2026-08-30T08:39:29.467469+00:00.