Support · Ollama

Install and connect Ollama

Ollama runs a small AI model on your own computer. It is free and works offline. In Cyer, Ollama is an optional backup — used when Claude or Codex cannot reply, or if you prefer not to use cloud AI.

1. 1

   Download and install Ollama

   Go to ollama.com/download and download the Windows installer. Run it and follow the prompts — same as installing any other app.

   You can also click Get Ollama inside Cyer’s setup window — it opens the same page.

2. 2

   Make sure Ollama is running

   After install, Ollama usually starts in the background. Look for a llama icon in the system tray (bottom-right of your screen, near the clock).

   If you do not see it, open Ollama from the Start menu, or open PowerShell and run ollama serve.

   

3. 3

   Open Cyer and start setup

   In Cyer, go to Settings → Integrations. Find Ollama (local fallback) and click Set up.

   The status badge at the top may say NOT INSTALLED, NOT RUNNING, NO MODELS, or READY. That tells you what to do next.

   

4. 4

   Pull a model

   Cyer’s default model is llama3.2. Click Pull llama3.2 in the setup window. The download runs in the background and can take several minutes (it is a large file).

   You can also pull manually in PowerShell:

   ollama pull llama3.2

   When the pull finishes, Cyer should detect the model and show it in the Model dropdown.

   Smaller models like phi3 download faster but may give shorter answers. If you use a different model, pull it first, then pick it in the Model list.
5. 5

   Select the model and test

   1. Choose your model from the Model dropdown (e.g. llama3.2).
   2. Click Test local AI. Cyer sends a short test message to Ollama.
   3. If it works, the badge should say READY and you will see a success message.

   The default Ollama address is http://127.0.0.1:11434. You only need to change this if you run Ollama on another computer (click Advanced — Ollama URL in the wizard).

   

6. ?

   Troubleshooting

   “Ollama is not running”	Start Ollama from the Start menu or run ollama serve. Click Refresh in Cyer.
   Model not found	Run ollama pull llama3.2 or pull the exact name shown in Settings.
   Very slow replies	Normal on CPU-only PCs. Try a smaller model like phi3.
   Cloud + Ollama both fail	Claude may be out of usage — make sure the Ollama tray icon is visible.

   To verify Ollama yourself, run in PowerShell: curl http://127.0.0.1:11434/api/tags — you should get JSON listing your models.

← All support guides