Free AI for your apps: four options and how to get their API key

Our apps (Tesora, Directorio, wpo-toolkit) come with their own engine for the heavy lifting. But for the recommendations and the text, they use an AI that lives in the cloud, not on your computer. That has a clear upside: you do not need a graphics card or an Ollama install. All it takes is a key from an AI, and several of them are free.

Here are four that work with our apps and how to get their key. Under the hood they all speak the same language (the OpenAI-compatible API), so in the app you only change three things: the URL, the model and the key. And the same key works for all of them.

What an API key is and why the app asks for one

An API key is a long password that tells the provider “this is me, bill the usage to my account”. The app stores it encrypted on your own computer and uses it to ask the AI for its opinion (which domain to buy, what text to write, and so on).

You will paste three things in Settings:

• AI URL: the provider’s server.
• Model: which brain to use.
• API key: your access password.

Let’s go through the four options.

1. Cerebras, the one we ship with

It is the fastest and the one we leave set by default. The free plan gives one million tokens a day with no card, on an 8K context. For normal use that is plenty.

1. Go to cloud.cerebras.ai and create the account. With Google or GitHub it is one click.
2. In the menu, open API Keys.
3. Click Create API Key, give it a name and copy it. It shows only once, so keep it safe.

In the app (Settings):

• URL: https://api.cerebras.ai/v1
• Model: gpt-oss-120b
• Key: the one you just copied

2. Google AI Studio, with Gemini for free

Google lets you use Gemini for free from AI Studio, and it has an OpenAI-compatible mode, which is exactly what our apps need.

1. Go to aistudio.google.com with your Google account.
2. Click Get API key, top left.
3. Create the key inside a project and copy it.

In the app (Settings):

• URL: https://generativelanguage.googleapis.com/v1beta/openai/
• Model: a flash model, whichever AI Studio shows you (for example gemini-2.0-flash)
• Key: the one you just copied

Google’s free plan has per-minute and per-day limits. For normal use that is usually enough.

3. OpenRouter, many models with a single key

OpenRouter is a gateway. With one key you reach models from several makers, and some are marked as free.

1. Go to openrouter.ai and create the account.
2. Open Keys in your profile and click Create Key.
3. Copy the key.

In the app (Settings):

• URL: https://openrouter.ai/api/v1
• Model: pick one whose id ends in :free (the free ones usually run at about 20 requests per minute)
• Key: the one you just copied

4. NVIDIA, with free credit to start

NVIDIA has its own model catalog with an OpenAI-compatible API and free credit when you sign up.

1. Go to build.nvidia.com and create the account. Signing up gives you free credit to start.
2. Pick a model from the list and click Get API Key.
3. Copy the key. It starts with nvapi-.

In the app (Settings):

• URL: https://integrate.api.nvidia.com/v1
• Model: the one you picked on the site
• Key: the one you just copied

Which one should I pick

If you do not want to think about it, leave Cerebras. It is the default and it has room to spare. The day it falls short or you want to try something else, you change those three things and you are done. Nothing stops you from holding accounts on several and switching between them.

Where you paste it in the app

1. Open the app (Tesora, Directorio or wpo-toolkit) and go to Settings.
2. Fill in AI URL, Model and API key with the details from the provider you chose.
3. Click Test connection. If it turns green, you are set: the recommendations now run on that AI.

It is the same screen and the same button in all our apps, so you learn it once and it works everywhere.

If you get an error, it is almost always the key pasted wrong (an extra space when copying) or the model spelled differently from how the provider names it. Check those two and try again.

That’s it

With this you have the AI working for you without spending a cent. Pick one, paste the three details, test the connection and get to work.

At SeoNova we build tools where the engine runs on your computer and the panel lives in the cloud. You can see the rest of what we do at seonova.pro.

— The SeoNova team