AI Application
Ollama — Local LLM Runtime
<h4>Run Llama, Mistral, Hermes & Phi locally — no API key needed</h4>
<p>Your own private AI inference server. Download any of 100+ open-source models (Llama 3.2, Mistral, Phi-3, Hermes, Gemma, Qwen, etc.) and call them via OpenAI-compatible API.</p>
<h5>What you get</h5>
<ul>
<li>Ollama latest with REST API on port 11434</li>
<li>Llama 3.2 1B pre-pulled (~1.3 GB) — runs fine on 4 vCPU CPU</li>
<li>Pull more models via SSH: <code>ollama pull llama3.2</code> · <code>ollama pull nous-hermes2:10.7b-solar-q4_K_M</code> · <code>ollama pull mistral</code></li>
<li>OpenAI-compatible endpoint at <code>POST /v1/chat/completions</code></li>
<li>1 TB outbound bandwidth / month</li>
</ul>
<h5>Performance note (CPU mode)</h5>
<p>Without GPU, expect ~5–30 tokens/sec on 7B models. Smaller models (1B–3B) are much faster and great for chat, summarization, classification.</p>
<h5>Tech specs</h5>
<p>8 vCPU · 8 GB RAM · 40 GB SSD · Ubuntu 24.04 LXC · ollama/ollama</p>
Plan
2
Configure
3
Review
4
Pay
Want to look at something else?
Switch to a different service — your cart will stay intact.
Back to all services