How to use Ollama on Windows as the LLM backend for Ozeki Voice Keyboard

This guide demonstrates how to configure Ollama as the LLM backend for Ozeki Voice Keyboard on Windows. By integrating a locally running Ollama instance, the AI assistant feature can process voice queries entirely on your machine: your speech is transcribed by the configured voice model, the resulting text is sent to Ollama as a prompt, and the generated response is automatically inserted into the active input field.

How it works

The diagram below illustrates the full pipeline of the AI assistant feature.

sequenceDiagram participant User participant OzKey as Ozeki Voice Keyboard participant Whisper as Voice Transcription Model participant Ollama as Ollama LLM (localhost) User->>OzKey: Voice input (audio) OzKey->>Whisper: Audio Whisper-->>OzKey: Return transcribed text OzKey->>Ollama: Send transcription as prompt Ollama-->>OzKey: Return generated response OzKey-->>User: Paste response into active field

Steps to follow

Before proceeding, make sure Ollama is installed on your Windows machine and at least one AI model has been downloaded and is ready to use. You can check out our guide about How to install Ollama

Verify Ollama is running
Open Ozeki Voice Keyboard and enable logging
Configure the LLM settings
Ask the AI assistant a question
Check the request in logs

How to use Ollama as the LLM backend video

The following video shows how to configure Ollama as the LLM backend for Ozeki Voice Keyboard step-by-step. The video covers configuring the LLM settings, using the AI assistant, and confirming the request in the log viewer.

Step 1 - Verify Ollama is running

Before configuring Ozeki Voice Keyboard, confirm that Ollama is running on your Windows machine and that the AI model you want to use is installed and available. Ollama exposes an OpenAI-compatible API on http://localhost:11434/v1 by default (Figure 1).

Running Ollama with installed AI model — Figure 1 - Ollama is running with an installed AI model

Step 2 - Open Ozeki Voice Keyboard and enable logging

Open Ozeki Voice Keyboard and locate its icon in the system tray in the bottom right corner of your taskbar (Figure 2).

Figure 2 - Open Ozeki Voice Keyboard

Before configuring the LLM settings, enable HTTP logging so you can verify that requests are reaching Ollama after setup. Right-click the tray icon and navigate to Logs from the context menu (Figure 3).

Navigate to logs from context menu — Figure 3 - Navigate to Logs from the context menu

In the Logs window, enable HTTP logging and close the window. Outgoing requests to Ollama will now be recorded and visible in the log viewer (Figure 4).

Enable HTTP logging and close window — Figure 4 - Enable HTTP logging and close the window

Step 3 - Configure the LLM settings

Right-click the tray icon and open the LLM settings from the context menu (Figure 5).

Open LLM settings from context menu — Figure 5 - Open LLM settings from the context menu

Enter the Ollama API URL and specify the model name you want to use. You can leave the API key field empty since a local Ollama instance does not require authentication. Click OK to save the settings (Figure 6).

http://localhost:11434/v1

Enter API URL, model and key — Figure 6 - Enter the Ollama API URL, model name and API key

Step 4 - Ask the AI assistant a question

Place your cursor in any input field where you want the response to appear, then press and hold Ctrl + Space and speak your question into the microphone. Release the keys when you have finished speaking. Your voice will first be transcribed, then the text will be forwarded to Ollama as a prompt (Figure 7).

Ask the LLM a question — Figure 7 - Ask the AI assistant a question using the hotkey

Once Ollama finishes generating the response, it is automatically pasted into the input field that was active when you started recording. No manual copying or pasting is required (Figure 8).

Figure 8 - The LLM response is automatically pasted into the active field

Step 5 - Check the request in logs

Open the Logs window to confirm that the request was sent correctly. You should see an HTTP request to the Ollama /v1/chat/completions endpoint, confirming that Ozeki Voice Keyboard is successfully communicating with the local Ollama instance (Figure 9).

View request in logs — Figure 9 - View the LLM request in the logs

Conclusion

You have successfully configured Ollama as the LLM backend for Ozeki Voice Keyboard on Windows. The AI assistant is now fully operational: press the hotkey, ask your question by voice, and the response will be generated by your local Ollama model and typed directly into whatever field is currently active on your screen.