Self-Hosting Overview

This section of the Colanode documentation explains how to deploy and run Colanode in your own infrastructure.

Self-hosting gives you complete control over your data, configuration, and scaling strategy. You can run Colanode for yourself, your team, or your organization, using a simple single-server setup or a production-grade distributed deployment.

Colanode is cloud-agnostic and works with most major hosting providers, bare-metal servers, or even local development environments. It’s designed to be flexible, so you can choose the infrastructure and tools that best fit your needs.

Dependencies

To self-host Colanode, you need PostgreSQL, Redis or Valkey, and an S3-compatible storage service.

Postgres

Colanode uses Postgres as its primary database to store structured information such as messages, documents, workspace metadata, user accounts, and permissions. Postgres can be hosted locally with Docker, installed directly on your servers, or provided by a managed cloud service like AWS RDS, Azure Database for Postgres, or Google Cloud SQL.

For Colanode’s AI-powered features, such as semantic search and embeddings, the database must have the pgvector extension enabled. This extension allows efficient storage and retrieval of vector embeddings generated by Colanode.

Redis or Valkey

An in-memory key-value store is required for caching, real-time event handling, and coordinating background tasks. Colanode supports both Redis and Valkey. These services can be deployed locally using Docker, installed directly on your servers, or hosted via a managed service such as AWS ElastiCache, Azure Cache for Redis, or Upstash.

S3-Compatible Storage

Colanode requires S3-compatible object storage for all unstructured content, including file uploads, images, and attachments. This can be any service that implements the S3 API, such as AWS S3, MinIO, Cloudflare R2 or Backblaze B2.

In the future we are planning to support other storage providers, such as Google Cloud or Azure storage services.