How to Get an API Key

Visiolog operates 100% in your browser. To transcribe paper registries and logbooks into spreadsheets, connect your preferred vision provider. Follow the step-by-step guides below.

Google Gemini API Key

Best transcription accuracy for handwritten logbooks & receipts

Open AI Studio

Visit Google AI Studio

Go to aistudio.google.com/app/apikey and sign in with any standard Google account.

Click "Create API key"

Click the blue Create API key button in the top right. Select an existing Google Cloud project or choose Create API key in new project.

Google AI Studio > Get API key > [ + Create API key ]

Copy your Secret Key

A string starting with AIzaSy... will be generated. Click the copy icon.

Key: AIzaSyD9x8K... (39 characters) [Copy]

Paste Key into Visiolog Settings

Return to Visiolog Settings, select Google Gemini, paste your key into the API key field, and click Test.

Free Tier Quota: Google AI Studio provides a free tier with 15 requests per minute, which is more than enough for daily paper-to-spreadsheet conversions.

Caution: Do not confuse Google Cloud OAuth Client IDs (ending in .apps.googleusercontent.com) with API keys. Visiolog expects a direct REST API key starting with AIzaSy....

Ready to configure?

Head over to settings and paste your key.

Open Settings

OpenRouter API Key

Access Claude 3.5 Sonnet, Llama 3.2 Vision, Qwen 2.5 VL, and more

Open OpenRouter

Navigate to openrouter.ai/keys and log in.

Create a New Key

Click Create Key. Enter a label (e.g. Visiolog Workspace) and click Create.

OpenRouter > Keys > [ + Create Key ]

Copy your Secret Key

Copy the generated key starting with sk-or-v1-....

Free Vision Models: OpenRouter hosts free vision models like meta-llama/llama-3.2-11b-vision-instruct:free.

Have your OpenRouter key?

Switch provider to OpenRouter in settings.

Open Settings

Ollama (100% Local Offline Vision)

Run vision AI on your own computer. Zero cloud fees, zero data leaves your machine.

Download Ollama

What is Ollama?

Ollama is a free program that runs artificial intelligence directly on your personal computer (using your GPU or CPU). Because it runs locally, you do not need an API key or an internet connection to transcribe confidential paper records, visitor books, or financial ledgers.

Download & Install Ollama

Go to ollama.com/download. Download the installer for your operating system (Windows, macOS, or Linux) and run it just like any regular desktop app.

Download a Vision AI Model

Open your Terminal (macOS/Linux) or PowerShell / Command Prompt (Windows), type the command below, and press Enter:

ollama run llama3.2-vision

What happens next: Ollama will automatically download the Llama 3.2 Vision model (~7.9 GB). Once download finishes, you can type /bye to exit the chat; Ollama remains running in the background as a local service.

Lightweight Alternative: If you have limited disk space or RAM, run ollama run minicpm-v (~5.5 GB), which also provides excellent document table OCR.

Enable Browser Access (One-Time CORS Setup)

Because Visiolog runs securely in your web browser, modern web security requires Ollama to allow incoming requests from websites. Choose your operating system below:

Method A (Fastest via PowerShell): Open PowerShell and run:

[System.Environment]::SetEnvironmentVariable("OLLAMA_ORIGINS", "*", "User")

Method B (Visual Windows Settings):

Open the Windows Start Menu, type Environment Variables, and click Edit the system environment variables.
Click the Environment Variables... button at the bottom.
Under User variables, click New...
Set Variable name to OLLAMA_ORIGINS and Variable value to *.
Click OK on all dialogs.

Crucial Step: Right-click the Ollama icon in the Windows taskbar system tray (bottom-right near clock), click Quit, then launch Ollama again from the Start Menu to apply changes.

Open the Terminal app on your Mac and run this command:

launchctl setenv OLLAMA_ORIGINS "*"

Crucial Step: Click the Ollama icon in your top Mac menu bar, choose Quit, and re-open the Ollama application from Applications.

If Ollama is installed as a systemd service, update the service override:

sudo systemctl edit ollama.service

Add these two lines under the editor:

[Service] Environment="OLLAMA_ORIGINS=*"

Save, exit, and reload the service:

sudo systemctl daemon-reload && sudo systemctl restart ollama

Verify Ollama is Running

Click this test link to open your local endpoint: http://localhost:11434.

Browser: http://localhost:11434 "Ollama is running"

Connect in Visiolog Settings

Go to Visiolog Settings, click the Ollama (Local) tab, and click Test. You are ready to transcribe!

Frequently Asked Questions

Is Ollama completely free?

Yes. Ollama is 100% open-source and free forever. There are no credit cards, subscriptions, or token charges.

Do I need a dedicated GPU?

An NVIDIA GPU or Apple Silicon Mac (M1/M2/M3/M4) delivers the fastest transcription speeds (2-5 seconds per page). However, Ollama will also run on standard Intel/AMD CPUs (takes ~15-30 seconds per page).

Why do I get "Failed to fetch" during the Test button?

This almost always means either (1) Ollama is not currently running, or (2) the OLLAMA_ORIGINS="*" environment variable was set but Ollama was not restarted afterwards. Quit Ollama from your taskbar or menu bar and re-launch it.

Ollama running locally?

Head over to Settings and verify your connection.

Open Settings

Custom OpenAI-Compatible Endpoint

Connect vLLM, LM Studio, Groq, Azure OpenAI, or self-hosted proxies

Locate your Base URL & Key

Ensure your endpoint supports OpenAI-compatible /v1/chat/completions with base64 vision image payloads.

Enter URL & Key in Settings

In Settings, select Custom Endpoint. Enter your custom server URL (e.g. https://api.groq.com/openai/v1/chat/completions), your API key, and your vision model name.

Have custom credentials?

Configure your custom endpoint in settings.

Open Settings