# open-webui/pipelines

Pipelines: Versatile, UI-Agnostic OpenAI-Compatible Plugin Framework

Repository: https://github.com/open-webui/pipelines
Canonical: https://ross.abutalabs.com/products/open-webui-pipelines
Language: Python
License: MIT
License Family: permissive
Topics: open-webui
Last push: 2025-08-18T23:05:01+00:00

## Health v2 (maintenance only)
Score: 41/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 37, release rhythm 35, longevity 59
- inputs: {"age_days": 834, "days_push": 380, "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 2428, forks 687 (observed 2026-08-28T04:06:50.948726+00:00)

## What it is
Pipelines is a UI-agnostic, OpenAI API-compatible plugin framework for adding custom Python logic and workflows to any client that speaks the OpenAI API spec. It is an Open WebUI initiative for offloading computationally heavy tasks like custom RAG, function calling, and message filtering from the main UI instance.

## Use cases
- add custom logic to an openai-compatible llm endpoint
- build a custom rag pipeline with llamaindex
- monitor llm messages with langfuse or opik
- add rate limiting to llm requests
- filter toxic messages in chat responses
- offload heavy model inference from open webui
- implement function calling filters for llm apps

## When to choose
- you need to offload computationally heavy tasks from your Open WebUI instance
- you want custom Python pipelines compatible with any OpenAI API client
- you need advanced filters like rate limiting, translation, or toxicity detection

## When to avoid
- you just want to add providers like Anthropic or simple filters - use Open WebUI Functions instead
- your client does not support pipe-type pipelines and you need filter features
- you need a lightweight solution without running a separate service

## Facets
- artifact type: framework
- maturity: active
- function: middleware, api-framework, rag, agent-framework, rate-limiting, monitoring, llm-inference
- domain: large-language-models, backend, apis, developer-tools
- platform: python, self-hosted, cross-platform
- tags: openai-compatible, open-webui, plugin-framework, llm-pipelines, filter-pipelines, ai-agents, retrieval-augmented-generation, docker

## Member repositories
- open-webui/pipelines (main) score 41

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:06:50.948726+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 2026-08-30T02:31:33.577047+00:00, confidence not recorded.
  - readme: https://github.com/open-webui/pipelines (fetched 2026-08-28T04:06:50.948726+00:00, sha 87c1bf9640e7)
- Data as of 2026-08-30T08:39:29.467469+00:00.