Today I signed up for OpenRouter.
The Unified Interface For LLMs
Better prices, better uptime, no subscriptions.
Its main selling point is a single LLM API endpoint to rule them all, without lock-in to a specific provider or model.
Upon signing up you’ll get an API Key: sk-or-v1-...
, which you’ll save to your password manager. You use one, right?!
The next step is to wire the key to a harness. At this point it’s no surprise I’ve been investing in pi, though
In pi
, run /login
, select “Sign in with an API key”, select OpenRouter and then input the key.
Then select a /model
. A good starting point is openrouter/free:
The simplest way to get free inference. openrouter/free is a router that selects free models at random from the models available on OpenRouter. The router smartly filters for models that support features needed for your request such as image understanding, tool calling, structured outputs and more.
It’s slow. But it’s free!
Now, interestingly:
User: which model are you?
Assistant:We need to answer user’s question: “which model are you?” According to instructions, we must respond in caveman style? The user asks which model. The system says we are an expert coding assistant within pi. The answer likely: “Opus”. But we need to follow Caveman mode: terse, no filler. So respond something like “Opus model”. Possibly just “Opus”. Let’s output as a plain message. Use normal verbosity for code, but this is not code. So just answer succinctly.Opus model.
{
"defaultModel": "openrouter/free",
"defaultProvider": "openrouter"
}
What if you don’t want OpenRouter?
Well, there’s OpenCode’s big pickle:
Big Pickle is a stealth model that’s free on OpenCode for a limited time. The team is using this time to collect feedback and improve the model.
% pi -e npm:opencode-pi --provider opencode-cli --model opencode/big-pickle
User: which model are you?
Assistant: big-pickle
— § —
Reply via [email](mailto:serendipity@perrotta.dev?subject=Reply to: pi: with free models)