AI/TLDR

PipesHub

Self-hosted enterprise search and agent context layer with permission-aware answers and block citations

RAG Frameworks & PlatformsOpen source
Language
Python
License
Apache-2.0
$curl -fsSL https://get.pipeshub.com/install | bash

Overview

PipesHub is an open-source workplace AI platform that connects the knowledge scattered across a company's business systems to AI applications. It ships 50+ enterprise connectors — Slack, Google Drive, GitHub, Microsoft 365, Notion and others — with both real-time and scheduled indexing, and puts a single search and question-answering layer over all of them. The same index can be exposed to your own agents, workflows and MCP clients rather than only to PipesHub's own UI.

Two properties drive its design. Answers are explainable: responses carry precise block-level citations back to the source document, so a user can check the passage a claim came from. And retrieval is permission-aware: PipesHub enforces source-level access controls, so each person only sees results they were already authorised to see in the originating system. Retrieval itself is graph-backed, capturing relationships across enterprise data rather than treating every chunk independently.

The project is Apache-2.0 and designed to be self-hosted: you bring your own LLM provider and deploy inside your own VPC, so documents never leave your infrastructure. It runs as a Docker Compose stack driven by an interactive installer, and publishes Node.js, Python and Go SDKs plus an MCP package for programmatic access.

What it does

  • 50+ enterprise connectors with real-time and scheduled indexing out of the box
  • Permission-aware search that enforces source-level access controls per user
  • Grounded answers with precise block citations back to the original documents
  • Knowledge-graph-backed retrieval that captures relationships across enterprise data
  • Bring-your-own-model and fully self-hostable — any LLM provider, deployed in your own VPC
  • No-code visual agent builder, artifacts and code execution in a sandbox, plus APIs, SDKs and MCP tools

Getting started

PipesHub runs as a Docker Compose stack. You need Docker with Compose v2; the interactive installer generates secrets and a .env file, picks a slim or full deployment, pulls images and waits for the stack to become healthy.

Run the one-command installer

This downloads the deployment files for the latest release into ./pipeshub and launches the interactive installer. When it finishes, open http://localhost:3000.

bashbash
curl -fsSL https://get.pipeshub.com/install | bash

Or inspect the script before running it

If you would rather read the installer first, download it, review it, and then run it yourself.

bashbash
curl -fsSL https://get.pipeshub.com/install -o pipeshub-install.sh
less pipeshub-install.sh        # review it
bash pipeshub-install.sh

Install from a cloned repository

To build from source, contribute, or pin the installer to your own checkout, clone the repo and run the same installer from the repo root. Building local images from source requires this path (./install.sh --build); the one-command installer always uses prebuilt images.

bashbash
git clone https://github.com/pipeshub-ai/pipeshub-ai.git
cd pipeshub-ai
./install.sh

Useful installer flags

Pass -y or --yes to accept all defaults for CI, --version TAG to pin a specific image tag, --reconfigure to re-run the wizard and overwrite an existing .env, and --print-env-only to write .env and print the compose command without starting containers. If you deploy on a cloud server, terminate TLS in front of PipesHub — browsers block some requests over plain HTTP, which shows up as a white screen.

bashbash
./install.sh -y --version 0.7.0

Connect it to your code

Once the stack is up, connect your business systems from the dashboard, then reach the same context layer programmatically through the published SDKs or the MCP package.

bashbash
npm install @pipeshub-ai/sdk
pip install pipeshub-sdk
npm install @pipeshub-ai/mcp

Commands and code are distilled from the project's own documentation — always check the official repo for the latest.

When to use it

  • Giving a company one search box over Slack, Drive, GitHub, Microsoft 365 and Notion without copying documents into a third-party cloud
  • Answering employee questions with citations a reviewer can open and check, instead of an unsourced summary
  • Supplying your own agents and MCP clients with permission-filtered enterprise context through one API
  • Running enterprise RAG entirely inside your own VPC when data residency or confidentiality rules forbid a hosted vendor

How PipesHub compares

PipesHub alongside other open-source rag frameworks & platforms tools AI/TLDR tracks, ranked by GitHub stars.

ToolStarsWhat it does
Dify★ 157kAn open-source platform with a visual workflow builder for creating LLM and RAG applications without writing much code.
graphify★ 120kTurns a folder of code, docs, PDFs and images into a local knowledge graph with tree-sitter AST parsing and Leiden communities — queryable by agents over MCP, no vector store.
RAGFlow★ 91.1kA RAG engine built around deep document understanding that turns complex files into a grounded, citation-backed question-answering layer.
Context7★ 62.3kContext7 pulls current, version-specific documentation and code examples for any library and feeds them into your LLM, available as a CLI skill or an MCP server.
Pathway★ 62.3kA Python framework with a Rust streaming engine that keeps ETL, real-time analytics and RAG pipelines continuously up to date as source data changes.
LightRAG★ 39.8kA graph-based RAG system that builds an entity-and-relationship knowledge graph for fast retrieval and easy incremental updates.
Quivr★ 39.5kQuivr is an open-source RAG framework that ingests your documents and answers questions about them, working with any LLM and any file type.
PipesHub★ 3.8kSelf-hosted enterprise search and agent context layer with permission-aware answers and block citations