PipesHub
PipesHub is an open-source context layer for AI agents that turns company knowledge from connected business systems into permission-aware search and cited answers, with self-hosted deployment and SDK/MCP access.
Visit pipeshub-ai/pipeshub-aiOverview
PipesHub is an open-source context layer for AI agents. It connects business systems such as Slack, Microsoft 365, GitHub, Google Drive, and Notion, then provides permission-aware search and trusted context with citations back to source documents. It combines a knowledge graph, page ranking, connectors, agents, MCP, and SDKs for building on company knowledge while keeping deployment under your control.
Key Features
- Permission-aware search across connected business systems, inheriting original document permissions so users only find content they can access.
- Verified citations with page numbers, paragraph references, and direct links to source documents across connected systems.
- Knowledge Graph and Page Ranking retrieval for structured reasoning across documents, entities, and metadata.
- Connectors to Slack, Microsoft 365, GitHub, Google Drive, Notion, Jira, Salesforce, and other business systems.
- No-code agent builder plus Python, Go, and TypeScript SDKs, with an MCP server to expose the knowledge graph to tools such as Claude, Cursor, and VS Code.
- Model-agnostic configuration with OpenAI, Anthropic, Google, Mistral, Ollama, or local LLMs.
Use Cases
- Answer employee questions over company knowledge with citations back to source documents.
- Search across Drive, Slack, Jira, email, and other connected systems from one permission-aware workspace.
- Build department-specific AI agents and event-driven or scheduled workflows without code, or with Python, Go, and TypeScript.
- Expose a company knowledge graph to external AI tools and developer products through MCP and SDKs.
Getting Started
- Try the hosted product for free, or deploy the open-source edition on your own instance.
- Start with the open-source edition or contact the team for enterprise deployment.
Deployment & Requirements
- Deployment options include cloud, VPC, on-premise, and air-gapped environments.
- The platform can scale from a single Docker deployment to enterprise-wide infrastructure.
- Model access can use your own keys from OpenAI, Anthropic, Google, or Mistral, or run fully local with Ollama.
Before You Adopt
- License: Apache-2.0. Review its terms before using, modifying, or distributing the project.