Oh My Pi · Laras Router Docs

This guide connects Oh My Pi to Laras Router. Every step was checked against the Oh My Pi repository.

You need an API key (lr_live_…) and a topped-up balance. New to Laras Router? Start with the Quickstart.

Set up

Install Oh My Pi

Install with Bun 1.4.4 or newer. The command is omp.

Shell

bun install -g @oh-my-pi/pi-coding-agent

Or use the shell installer: curl -fsSL https://omp.sh/install | sh.

Set your API key

Put your key in an environment variable.

Shell

export LARAS_API_KEY="lr_live_your_key_here"

Add the provider

Create ~/.omp/agent/models.yml.

~/.omp/agent/models.yml

providers:
  laras:
    baseUrl: https://api.larasrouter.com/v1
    api: openai-completions
    apiKey: LARAS_API_KEY
    models:
      - id: deepseek-flash
        name: DeepSeek-V4.1-Flash
        reasoning: true
        input: [text]
        contextWindow: 1000000
        maxTokens: 8192

Pick the model

Start omp, run /model, and choose DeepSeek-V4.1-Flash from the laras provider. To start with the model already selected, run omp --model laras/deepseek-flash.

Good to know

apiKey names an environment variable. Write LARAS_API_KEY, not the key itself. If the variable is unset or empty, Oh My Pi uses the text LARAS_API_KEY as the token, so export it before you start omp.
Oh My Pi reads models.yml before models.yaml. Both live in ~/.omp/agent.
contextWindow states the full 1,000,000-token window. maxTokens is the output cap, and it also sets how much balance a request must hold while it runs. 8,192 is the default output cap, so raise it only if you need longer replies, and keep enough balance for it.
input lists text only. deepseek-flash does not accept image input.
Create a separate API key for Oh My Pi with a monthly budget, so you can see and cap what it spends.

Settings and screens can change between versions. If something looks different, check the Oh My Pi repository. For other tools, see all integrations.