How to Deploy a Lightweight Gitea Git Server with Docker for AI Platforms

Background and Motivation

When building an AI model‑hosting platform similar to ModelScope, each project needs an independent Git repository service. The initial choice of GitLab quickly hit a resource ceiling, consuming up to 10 GB of memory, which is unsustainable for a lightweight deployment.

Choosing Gitea

Switching to Gitea, a lightweight self‑hosted Git solution, reduced memory usage to around 600 MB and provided a full set of REST APIs, enabling deep integration with the AI platform.

Prerequisites

Database (MySQL / PostgreSQL / SQLite) – required for storing core data such as users, repositories, issues, PRs, permissions, and LFS metadata. The tutorial uses MySQL 8.

S3 / MinIO (Object Storage) – optional; used for repository attachments and large files, otherwise local disk is used.

Redis – optional; can serve as cache, session store, or background queue to improve performance.

Docker Installation

Run the following command to start Gitea in a Docker container:

docker run -d --name gitea \
  -p 3000:3000 -p 222:22 \
  -v /Users/zhangyong/docker/gitea/data:/data \
  -v /etc/localtime:/etc/localtime:ro \
  -v /etc/timezone:/etc/timezone:ro \
  -e USER_UID=1000 \
  -e USER_GID=1000 \
  --restart always \
  gitea/gitea:latest

Initial Setup

After the container starts, open http://localhost:3000. On the first‑run page, configure the database connection (e.g., MySQL 8) and click “Install”. The configuration file is written to /data/gitea/conf/app.ini. Once installed, the login screen appears.

Creating Repositories

The repository creation UI resembles GitHub. Enter a repository name and click “Create”. To clone or push, generate a Personal Access Token (PAT) for a user.

Using the REST API

Visit http://localhost:3000/api/swagger#/repository to explore all Gitea development APIs. These endpoints can be wrapped into custom services, such as a Java client library, to automate repository management.

Conclusion

The comparison chart shows Gitea’s lightweight footprint versus GitLab’s heavyweight requirements. Gitea proved to be a compact, feature‑rich solution that fits the company’s AI platform needs while keeping resource consumption low.