Back to catalog
Ollama logo

Self-Host with Docker · Yantr

Ollama

Run large language models locally - Llama, Mistral, CodeLlama, and hundreds more.

About

Ollama is a powerful platform for running large language models locally on your own hardware. It supports hundreds of models including Llama 2/3, Mistral, CodeLlama, Phi, and many more. Just download a model and start chatting or integrate via API. It handles model downloading, memory management, and inference optimization. Perfect for privacy-conscious AI enthusiasts who want to run their own AI without cloud dependencies. Models run entirely on your GPU (or CPU), and your conversations never leave your machine.

How to Self-Host Ollama

Ollama can be deployed with standard Docker Compose using the Yantr app catalog. This page collects the basic information searchers usually need first: what Ollama does, which Docker image Yantr uses, default ports, dependencies, and where to find the original compose template.

If you searched for Ollama Docker, Ollama Docker Compose, or self-host Ollama, this page is the direct landing page for that app in Yantr's catalog.

Use Cases

  • Run Llama 2, Mistral, or other open-source LLMs completely offline.
  • Build AI-powered applications that use local models via REST API.
  • Experiment with different AI models without API costs.
  • Create a private AI assistant for coding, writing, or analysis.

Technical Details

Ports

Port Protocol Label
11434 HTTP API

Docker Image

ollama/ollama:latest

Service Name

ollama

Notes

  • Requires adequate GPU (VRAM) or CPU for acceptable performance
  • Models downloaded on first use - expect several GB per model
  • API available at port 11434 for integration with other apps

Related Apps

More App Pages

Explore the full app catalog for more self-hosted services, alternatives, and dependencies that work alongside Ollama.

Page generated 2026-03-06T16:43:27.553Z