- Docker
19.0.3
or later. - Docker Compose (
docker compose
ordocker-compose
)2.6.1
or later. - OpenAI API key (for answer generation and embeddings). Get it from here.
- Firecrawl API key (for website scraping). Get it from here.
Run this command to install Gurubase:
curl -fsSL https://raw.githubusercontent.com/Gurubase/gurubase/refs/heads/master/gurubase.sh -o gurubase.sh
bash gurubase.sh
The installer will:
- Create a
.gurubase
directory in your home folder - Prompt for required API keys
- Download and start all necessary services
- Open the web interface at http://localhost:8029
You can upgrade to the latest version by running the following command:
curl -sSL https://raw.githubusercontent.com/Gurubase/gurubase/refs/heads/master/gurubase.sh -o gurubase.sh
bash gurubase.sh upgrade
You can remove Gurubase by running the following command. This will remove the containers and networks but keep the data files.
curl -sSL https://raw.githubusercontent.com/Gurubase/gurubase/refs/heads/master/gurubase.sh -o gurubase.sh
bash gurubase.sh rm
Caution
To remove everything including all data (volumes), you can run the following command:
rm -rf ~/.gurubase
-
Operating System
- macOS 10.15 Catalina or later
- Linux (Ubuntu 20.04 LTS, Debian 10, CentOS 8 or later)
⚠️ Native Windows is not supported, but you can use WSL2 to run Gurubase on Windows.
-
Processor
- Quad-core CPU (4 cores) at 2.5 GHz or higher
- ARM-based processors supported (e.g., Apple M1, M2, M3, etc.)
-
Memory and Storage
- 8GB RAM minimum recommended
- 10GB available disk space (SSD preferred for better performance)
-
Network
- Ports
8028
and8029
must be available
- Ports
Note
Only Linux and MacOS are supported at the moment. Native Windows is not supported, but you can use WSL2 to run Gurubase on Windows.
The following services are installed as part of Gurubase:
Service | Version | Description |
---|---|---|
Milvus | v2.4.17 | Vector database for storing and searching embeddings |
RabbitMQ | 3.13.7 | Message broker for task queue management |
PostgreSQL | 16.3 | Main database for storing application data |
Redis | 7.2.6 | In-memory data store for caching, rate limiting, and session management |
Nginx | 1.23.3 | Web server for routing and serving static files |
Gurubase Backend | latest | Django based backend application |
Gurubase Frontend | latest | Next.js based frontend application |
Celery Workers | latest | Background task processors (2 workers) |
Celery Beat | latest | Periodic task scheduler |
All data is stored locally in ~/.gurubase/
:
- PostgreSQL data:
~/.gurubase/postgres/
- Milvus data:
~/.gurubase/milvus/
- Redis data:
~/.gurubase/redis/
- Media files:
~/.gurubase/backend_media/
- Environment variables:
~/.gurubase/.env
You can access the Django admin interface with the following credentials:
URL | Password | |
---|---|---|
http://localhost:8028/admin |
root@gurubase.io |
ChangeMe |
After logging in, you can change the admin password from http://localhost:8028/admin/password_change/
.
Warning
This interface is intended for advanced users only. Be cautious when making changes as they can affect your Gurubase installation.
By default, Gurubase collects anonymous telemetry data using PostHog to help improve the product. You can opt out of telemetry by following these steps:
- Edit the frontend environment file
~/.gurubase/.env.frontend
:
NEXT_PUBLIC_TELEMETRY_ENABLED=false
- Recreate the frontend service to apply the new environment:
cd ~/.gurubase && docker compose up -d --force-recreate frontend
Here's a detailed comparison between Gurubase Cloud and Self-hosted versions:
Feature | Gurubase Cloud | Self-hosted |
---|---|---|
Authentication | ✅ Google & GitHub via Auth0 | ❌ Not available |
API Support | ✅ Full API access | ✅ Full API access |
Binge | ✅ Follow-up questions & learning graph | ✅ Follow-up questions & learning graph |
Knowledge Base Sources | ✅ Websites, YouTube, PDFs | ✅ Websites, YouTube, PDFs |
GitHub Codebase Indexing | ✅ Available | ✅ Available |
Website Widget | ✅ Available | ✅ Available |
Base LLM | ✅ OpenAI GPT-4o | ✅ OpenAI GPT-4o |
For frequently asked questions about Gurubase, including system architecture, use cases, data handling, and more, please check the FAQ section in README.md.