Ollama
Image: ollama/ollama:latest
- Port 11434/tcp
- 30 GB volume at /root/.ollama
- 4 vCPU · 8 GB
Template · AI & LLM
Ollama
Run open LLMs (Llama, Mistral, Gemma) with an OpenAI-compatible API.
Deploy on Skiffly
Opens the dashboard with this template pre-selected. Sign in with GitHub if you are new; the free tier needs no card.
- Ollamaollama/ollama:latest
What gets deployed
One service in the environment you pick, on the private network of the project.
About this template
ollama/ollama:latest, CPU inference, models cached on a volume. Pull a model with ollama pull llama3.2 via the API.
Variables
Generated secrets never leave the platform; references are resolved on the private network. Anything marked required is asked for at deploy time.
Ollama
| Name | Value |
|---|---|
| OLLAMA_HOST | 0.0.0.0:11434 |
| OLLAMA_KEEP_ALIVE | 24h |
| OLLAMA_BASE_URL | http://${{self.SKIFFLY_PRIVATE_DOMAIN}}:11434 |
Deploy button for your README
Add a one-click deploy link to your repository or docs. Two variants: for light and dark backgrounds.
Markdown · Light background
[](https://skiffly.dev/templates/ollama)Markdown · Dark background
[](https://skiffly.dev/templates/ollama)HTML
<a href="https://skiffly.dev/templates/ollama"><img src="https://skiffly.dev/button.svg" alt="Deploy on Skiffly" height="32"></a>Related templates
All templates- Open WebUI + OllamaChatGPT-style UI for local models, wired to an Ollama service.
- AnythingLLMAll-in-one RAG chat app with built-in vector DB and document ingestion.
- FlowiseDrag-and-drop builder for LLM agents and RAG flows, PostgreSQL-backed.
- LangfuseLLM observability and prompt management (v3 stack: Postgres, ClickHouse, Redis, MinIO).
Deploy Ollama in one click
Volumes, secrets and domain pre-wired. Runs on EU nodes, paid per minute from a prepaid balance.