Setup

Get an OpenRouter Key for Free Models

Want your agent to use free models and try a huge menu of options through one API key? This guide shows you how to get an OpenRouter key, add it to heyron, and pick a few free models worth trying first.

⏱ 6 minute setup

Why OpenRouter?

OpenRouter gives you one API key that can access 300+ models through a single service. That includes paid models and a rotating set of free models.

For a heyron user, that means:

One key, lots of models — you don't need separate accounts for every provider just to experiment.
Free options exist — great for learning, testing, and low-volume usage.
Easy model switching — if one free model feels weak, you can try another fast.

Important context

You do not need this to start using heyron. Your heyron subscription already includes AI usage. This tutorial is for people who want their own OpenRouter key so they can experiment with OpenRouter's free models and wider model catalog.

What You'll End Up With

By the end

You'll have an OpenRouter API key, you'll know where to paste it in heyron, and you'll have a short list of free models to try without getting lost in a giant model menu.

Step 1: Create an OpenRouter Account

Sign up

Go to openrouter.ai and create an account.

If you're already signed in, great — move to the next step.

Step 2: Generate Your API Key

Make the key

Open the API keys page on OpenRouter and create a new key.

Direct link: openrouter.ai/keys

Your key will look like a long secret string. Copy it somewhere safe right away.

Treat it like a password

Do not post your API key in Discord, Telegram, screenshots, GitHub, or any public channel. If you accidentally leak it, revoke it and create a new one.

Step 3: Add It to Your Agent

Paste the key into heyron

Open your heyron dashboard and add your OpenRouter key in your agent's model/API key settings.

If you prefer, you can also tell your agent what you want in plain English, like:

Use my OpenRouter API key for model access.

If you're not sure where to paste it

Use the dashboard's model/provider settings area. If heyron's UI wording changes, the idea stays the same: add your OpenRouter key as a provider key, then choose an OpenRouter model.

Step 4: Pick a Free Model

OpenRouter has a lot of models. Do not make this harder than it needs to be.

You have two easy options:

Option A: use the free router — OpenRouter supports openrouter/free, which automatically chooses an available free model.
Option B: choose a specific free model — look for models marked (free) in OpenRouter's list.

Simple recommendation

If you are new, start with the free router or one of the suggested models below. Don't spend 45 minutes comparing model names before you even start using your agent.

Free Models Worth Trying First

Free models change over time, but these are solid types of picks to try first on OpenRouter:

OpenRouter Free Router

Best for: beginners. Lets OpenRouter choose an available free model for you. Lowest-friction place to start.

DeepSeek R1 (free)

Best for: reasoning, problem-solving, step-by-step thinking. Good when you want the model to "think through" something.

Trinity Large Preview (free)

Best for: general chat, experimentation, trying a stronger open-weight model without paying.

Trinity Mini (free)

Best for: faster, lighter everyday use when you want speed over depth.

Llama models (free)

Best for: general purpose chatting and trying popular open models. Good backup option if another free model is busy or weak.

Any model marked “(free)”

Best for: experimenting. If one free model feels weird, slow, or weak, switch. That is normal.

What to Expect From Free Models

Free models are great for learning and light use, but set expectations correctly:

They may be slower than paid models.
They may have lower limits or temporary availability issues.
Quality can vary more from one model to another.
They are perfect for experimenting, not always ideal for your most important production work.

Free model rate limits

OpenRouter's docs say free model access is rate-limited based on how many credits you've purchased. If you've purchased less than 10 credits, you're limited to 50 free-model requests per day. If you've purchased at least 10 credits, that increases to 1000 free-model requests per day. In other words: a one-time $10 credit purchase can unlock a much more usable daily free-model limit, and you do not need to spend those credits just to use the free requests.

Do not overthink the first choice

The goal is not to find the perfect free model on day one. The goal is to get your agent chatting successfully, then swap models if needed.

Best Beginner Setup

If you want the easiest path, do this:

Use OpenRouter for variety.
Start with the free router or a clearly marked free model.
Use your agent for real tasks for 10–15 minutes.
If it feels weak, switch models instead of assuming your agent is broken.

Troubleshooting

“My OpenRouter key doesn't work”

Double-check that you copied the whole key and pasted it correctly. If needed, create a new key and try again.

“I added the key but can't tell which model to use”

Start with the free router or any model marked (free). Don't wait for a perfect choice.

“The free model is bad / weird / slow”

That's not unusual. Switch to another free model. Free models are for experimentation, so some will feel better than others.

“Can I stay free forever?”

Maybe for light use. But free models can change, rotate, or have limitations. If you start depending on your agent heavily, a paid model may eventually make more sense.

The Short Version

If you only remember 3 things

1. OpenRouter gives you one key for 300+ models.
2. Some of those are free.
3. Start simple: use the free router or a model marked (free) and begin chatting.

What's Next?

🧠 Understand providers and model choices — API Keys & Models →

💬 Learn how to prompt your agent better — Prompting Your Agent Like a Pro →

📁 Learn what to save so good work doesn't disappear — How Agent Memory Works →