What it is

PipesHub is an open-source workplace AI platform in the enterprise AI ecosystem. It connects AI applications to knowledge stored across business systems, including Slack, Google Drive, GitHub, Microsoft 365, Notion, and 50+ other systems. The project gives AI agents trusted context and gives teams permission-aware search with verified citations across those sources.

The problem it solves is the gap between scattered enterprise knowledge and reliable AI answers. Scattered documents, conversations, code, and files often leave users and agents without source context or access-control enforcement. PipesHub indexes connected content, retrieves it through search and knowledge-graph methods, and filters results according to source-level permissions so each person sees only authorized material.

Key capabilities

• Permission-aware search enforces source-level access controls, so users retrieve only authorized documents and records.
• Explainable answers provide grounded responses with precise block citations to original documents.
• Enterprise connectors support real-time and scheduled indexing across Slack, Google Drive, GitHub, Microsoft 365, Notion, and other systems.
• Knowledge-graph retrieval uses graph-backed search to capture relationships across enterprise data.
• Multimodal support covers images, diagrams, scanned files, and voice-based interaction; audio and video are coming soon.
• Agents and workflows use the same context layer through APIs, SDKs, MCP tools, custom connectors, and no-code agent building.
• Self-hosting with bring-your-own-model support allows deployment in a private VPC with any LLM provider, keeping data inside customer infrastructure.

Who uses it and how

• Enterprise teams use it as a permission-aware search layer across internal documents, conversations, and repositories.
• AI application developers use it to supply trusted context to agents, workflows, and MCP clients.
• Operations teams use it to build no-code agents and actions that execute across connected enterprise tools.
• Analysts and engineers use it to generate reports, charts, and dashboards inside a safe execution sandbox.
• Self-hosters deploy it locally or in their own infrastructure to keep enterprise knowledge and model traffic under their control.

Getting started

The README recommends a single-command installation using curl -fsSL https://get.pipeshub.com/install | bash, and it states that PipesHub can be run locally or deployed. A managed PipesHub Cloud option is described as coming soon, with a waitlist link.

When to use it — and when not to

PipesHub fits organizations needing self-hosted, permission-aware enterprise search and agent context; its topic list includes Glean, indicating comparison to that product. Self-hosting requires operating a broad stack, including Neo4j or ArangoDB, Qdrant or OpenSearch or Redis, MongoDB, local filesystem or S3 or Azure Blob, Kafka or Redis Streams, Celery, FastAPI, LangChain, and document parsing or conversion tools. Repository data shows a young project, zero contributors, and 172 open issues, so teams needing a mature, low-operations replacement for enterprise search products should evaluate maintenance risk.

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