LowRouterBeta

LowRouter documentation

LowRouter is an HTTP gateway for large language models. It speaks the OpenAI API on the front, fans out to multiple providers on the back, and attaches energy and carbon estimates to every response. You get one endpoint, one credit balance and one usage history, and every call comes with a footprint estimate.

These docs are written for developers and operators. Every page is plain markdown: append .md to any URL or use the Copy page button to get the raw source.

What’s available today

LowRouter is in an invite-only beta, run as an early-stage, experimental service without an SLA. In practice:

  • Text inference is live: chat and text completions through the OpenAI-compatible API and the Anthropic-native /v1/messages, across the providers in the model catalog.
  • Image input is live: images, and documents and audio where the provider accepts them, as content blocks on either surface. Which routes take which is per provider; see input modalities.
  • Image and video generation, and MCP, are not available yet. They’re on the roadmap and listed as “coming soon”.
  • Signup requires an invite link for now.

If a capability isn’t on the “live” list above, the API doesn’t support it today.

Where to start

If you have ten minutes:

  1. Create an account
  2. Create your first API key
  3. Run your first completion

If you want to understand the project before you sign up, the Philosophy section explains what LowRouter is, what it isn’t, and why we built it.

What’s in this site

  • Philosophy: the principles, the position, and what we explicitly do not claim.
  • Getting started: account setup, keys, first request, dashboard tour.
  • Guides: day-to-day operations such as keys, credits, billing, usage, and reading the dashboard.
  • Integrations: copy-paste snippets for the OpenAI SDK and common LLM developer tools.
  • Models & providers: what’s routable, how routing works, and how to read per-request metadata.
  • Sustainable AI: how we estimate energy and carbon, what data we use, and what the numbers mean.
  • FAQ: short answers to the questions we get most often.

API at a glance

Bash
curl https://api.lowrouter.ai/v1/chat/completions \
  -H "Authorization: Bearer $LOWROUTER_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "auto/mistralai/mistral-large-2512",
    "messages": [{"role": "user", "content": "Hello"}]
  }'

The response carries the usual OpenAI-shaped JSON plus a lowrouter_metadata block with the resolved route and the energy and carbon estimate for the request. See per-request metadata for the schema.

Source and licence

The platform is operated by Carbonifer SAS. Source for these docs lives in the docs-public/ directory of the LowRouter repository. Page content is published under the same terms as the rest of the site; see the legal pages linked in the footer.