Use Ollama with Pluely: Fully Local, Fully Offline AI

Connect Ollama and the Pluely overlay answers with models running entirely on your machine — nothing leaves your computer, and it works offline.

1. Install and pull a model

Install Ollama from ollama.com, start it, and pull at least one model in your terminal — for example ollama pull llama3.2 (or gemma3 for vision, deepseek-r1 for reasoning).

2. Connect it in Pluely

1. Open the dashboard → Dev space → Connect your CLI.
2. Pluely detects Ollama automatically. Click Connect.
3. Pick a model — required for Ollama. The Model panel lists everything you've pulled locally; the list refreshes when you reconnect.
4. It appears at the top of the overlay's model picker as Your CLI (e.g. Ollama · llama3.2).

Vision models

Attach screenshots and images when your selected model is multimodal — Pluely recognizes the vision families (llava, gemma3, qwen-vl, llama3.2-vision, minicpm-v, and more) and enables the image controls automatically. Text-only models keep them disabled with a hint to pull a vision model.

Thinking models

Reasoning models like DeepSeek-R1 and Qwen3 are fully supported: their chain-of-thought is stripped automatically, so the overlay shows only the final answer — never a wall of <think> text.

Troubleshooting

• "Ollama isn't running" — start the Ollama app (or ollama serve) and hit the refresh button beside "Connect your CLI".
• No models in the picker — ollama pull one in your terminal, then reconnect.

Related

• Connect a local CLI (all options)
• Local-first & privacy
• Choosing your AI