Deploying Onyx Community Edition

Introduction

Onyx Community Edition (formerly Danswer) is an open-source AI-powered enterprise search and chat assistant that connects to your organization’s documents and knowledge bases. It enables teams to ask questions in natural language and receive accurate answers sourced from internal documentation, Slack conversations, Confluence pages, GitHub repositories, and many other data sources.

Built with Python, TypeScript, and leveraging modern LLM technology, Onyx provides a ChatGPT-like experience tailored to your organization’s specific knowledge. The platform includes built-in connectors for popular enterprise tools and supports custom connector development.

Key features of Onyx include:

Natural Language Search: Ask questions in plain English and get contextual answers
Multi-Source Connectors: Connect to Slack, Confluence, Google Drive, GitHub, Notion, and 30+ other sources
Citation and Sources: Every answer includes citations to original source documents
Permission Aware: Respects existing document permissions from connected sources
Chat Interface: Conversational AI assistant for follow-up questions
Custom Personas: Create specialized assistants for different teams or topics
Self-Hosted LLM Support: Use OpenAI, Anthropic, or self-hosted models
Enterprise Ready: SSO, audit logs, and team management features

This guide walks through deploying Onyx Community Edition on Klutch.sh using Docker.

Why Deploy Onyx on Klutch.sh

Deploying Onyx on Klutch.sh provides organizations with a private AI assistant:

Data Privacy: Keep your organization’s documents and search queries on infrastructure you control rather than sending them to third-party services.

Persistent Storage: Store vector embeddings and indexed documents with persistent volumes that survive updates.

Scalable Resources: Allocate CPU and memory based on document volume and query load.

HTTPS by Default: Secure access to your AI assistant with automatic SSL certificates.

Custom Domains: Provide a memorable URL for your team to access the search assistant.

Prerequisites

Before deploying Onyx on Klutch.sh, ensure you have:

A Klutch.sh account
A GitHub account with a repository for your deployment
An LLM API key (OpenAI, Anthropic, or compatible provider)
A PostgreSQL database instance
Basic familiarity with Docker and Python applications

Deploying Onyx on Klutch.sh

Set Up Required Services

Onyx requires PostgreSQL and a vector database (Vespa or Qdrant). Deploy these services on Klutch.sh or use external instances.

Create Your Dockerfile

Create a Dockerfile in your repository:

FROM python:3.11-slim

# Install system dependencies
RUN apt-get update && apt-get install -y \
    build-essential \
    curl \
    git \
    && rm -rf /var/lib/apt/lists/*

# Clone Onyx repository
WORKDIR /app
RUN git clone https://github.com/onyx-dot-app/onyx.git .

# Install Python dependencies
WORKDIR /app/backend
RUN pip install --no-cache-dir -r requirements.txt

# Build frontend
WORKDIR /app/web
RUN curl -fsSL https://deb.nodesource.com/setup_18.x | bash - \
    && apt-get install -y nodejs \
    && npm install && npm run build

WORKDIR /app

EXPOSE 3000 8080

COPY start.sh /start.sh
RUN chmod +x /start.sh

CMD ["/start.sh"]

Create Startup Script

Create start.sh to start both backend and frontend:

#!/bin/bash
cd /app/backend
python -m uvicorn danswer.main:app --host 0.0.0.0 --port 8080 &

cd /app/web
npm start

Push to GitHub

Commit and push your Dockerfile and scripts to your GitHub repository.

Create a New Project on Klutch.sh

Navigate to the Klutch.sh dashboard and create a new project.

Create and Configure the App

Create a new app and connect it to your GitHub repository.

Configure HTTP Traffic

Set the traffic type to HTTP with an internal port of 3000 for the web interface.

Set Environment Variables

Configure required environment variables:

Variable	Description
`POSTGRES_HOST`	PostgreSQL host address
`POSTGRES_USER`	Database username
`POSTGRES_PASSWORD`	Database password
`OPENAI_API_KEY`	OpenAI API key (or alternative LLM key)
`SECRET_KEY`	Application secret key

Attach Persistent Volumes

Add persistent storage:

Mount Path	Recommended Size	Purpose
`/app/data`	50+ GB	Document index and embeddings
`/app/storage`	10 GB	File storage

Deploy Your Application

Click Deploy to build and launch your Onyx instance.

Configure Connectors

Access your deployment and configure connectors to your organization’s data sources through the admin interface.

Deploying Onyx Community Edition

Introduction

Why Deploy Onyx on Klutch.sh

Prerequisites

Deploying Onyx on Klutch.sh

Additional Resources