OpenRouter

By OpenRouter

One OpenAI-compatible API to 300+ models across providers, with routing and fallbacks

Website ↗ Docs ↗

APIProprietary

#gateway #multi-provider #routing #api

Free tier Paid

Doc version 1.0 · Updated June 28, 2026

Related concepts

Harness Engineering Doc version 1.0 · Updated June 28, 2026 Wrapping an LLM in scaffolding — a control loop, tools, a sandbox, guardrails, and evaluation — so it behaves reliably in production.

Related tools

LiteLLM

Gateways & Routers

OpenAI

Foundation Model APIs

Overview

OpenRouter is a gateway that exposes 300+ models from many providers behind one OpenAI-compatible API and key.
It adds routing, automatic fallbacks, and unified billing, so an agent can reach Claude, GPT, Gemini, Llama, and more without juggling separate accounts.

flowchart LR
  app[your app] --> or[OpenRouter]
  or --> a[Anthropic]
  or --> o[OpenAI]
  or --> g[others]

The Code samples tab shows the LiteLLM route and the OpenAI SDK route — pick from the selector to compare.

When to use it

Reach for OpenRouter when you want a single key and endpoint across many providers — to compare models, add fallbacks, or avoid managing each account.

Example

Python · via LiteLLM

When to use

Route through LiteLLM to reach any OpenRouter model with the same call shape as every other provider.

How it flows

The openrouter/ prefix tells LiteLLM to send the request to OpenRouter, which forwards it to the chosen model.

Structure

flowchart LR
  app[your app] --> lite[LiteLLM]
  lite -->|openrouter/| or[OpenRouter]
  or --> models[(300+ models)]

from litellm import completion

# openrouter/<vendor>/<model> routes through OpenRouter with OPENROUTER_API_KEY.
resp = completion(
    model="openrouter/anthropic/claude-opus-4-8",
    messages=[{"role": "user", "content": "Summarize OpenRouter in one line."}],
)
print(resp.choices[0].message.content)

When to use

Use the OpenAI SDK directly when a tool already speaks OpenAI and you just want to redirect it.

How it flows

OpenRouter mirrors the OpenAI API, so only the base_url and key change.

Structure

flowchart LR
  app[your app] --> or[OpenRouter]
  or --> a[Anthropic]
  or --> o[OpenAI]
  or --> g[others]

import os
from openai import OpenAI

# OpenRouter is OpenAI-compatible — point the base_url at it.
client = OpenAI(
    base_url="https://openrouter.ai/api/v1",
    api_key=os.environ["OPENROUTER_API_KEY"],
)
resp = client.chat.completions.create(
    model="anthropic/claude-opus-4-8",
    messages=[{"role": "user", "content": "Summarize OpenRouter in one line."}],
)
print(resp.choices[0].message.content)

Example

OpenRouter — one key, many models (via LiteLLM)

A real, runnable mini-project. Download it and run with Docker.

View samples/openrouter_1 on GitHub ↗

June 28, 2026

OpenRouter — one key, many models (via LiteLLM)

A tiny OpenRouter script: it sends one prompt through litellm.completion() and prints the reply. OpenRouter is an OpenAI-compatible gateway that fronts hundreds of models behind a single API key — switch between Claude, GPT, Gemini, Llama, and more by changing only MODEL, never the code.

Configure

cd samples/openrouter_1
cp .env.sample .env
# edit .env: paste your OPENROUTER_API_KEY (and optionally change MODEL)

MODEL is a LiteLLM id of the form openrouter/<vendor>/<model>:

Vendor / model	`MODEL`
Anthropic Claude	`openrouter/anthropic/claude-opus-4-8`
OpenAI	`openrouter/openai/gpt-4o`
Google Gemini	`openrouter/google/gemini-2.5-flash`
Meta Llama	`openrouter/meta-llama/llama-3.3-70b-instruct`

Every model uses the same OPENROUTER_API_KEY. Get one at openrouter.ai/keys. Browse model ids at openrouter.ai/models. .env is gitignored — only .env.sample is committed.

Tip: model ids ending in :free (e.g. openrouter/openai/gpt-oss-120b:free) run without purchased credits, subject to rate limits. Paid models such as openrouter/openai/gpt-4o need credits on your account, or the request returns 402 Insufficient credits.

Run with Docker

cd samples/openrouter_1
docker build -t aas-openrouter .
docker run --rm --env-file .env aas-openrouter "Summarize OpenRouter in one line."

Run with Docker (in a devcontainer with DooD)

In a dev container that talks to the host Docker daemon (Docker-outside-of-Docker), the foreground docker run above often prints nothing and exits 0 — but the run itself succeeds. The script runs to completion and Docker captures all of its output; only the live attached stream drops it over the VM boundary. Run detached and follow the logs instead:

cd samples/openrouter_1
docker build -t aas-openrouter .
docker logs -f "$(docker run -d --env-file .env aas-openrouter \
  "Summarize OpenRouter in one line.")"

Run locally

cd samples/openrouter_1
pip install -r requirements.txt
python app.py "Summarize OpenRouter in one line."

python-dotenv loads .env automatically.

Example run

Output varies by model and run — LLMs are non-deterministic, so the exact wording differs each time. Below is one run with openrouter/openai/gpt-oss-120b:free.

OpenRouter is an AI gateway that aggregates multiple language model providers into a single, unified API for easy, cost‑effective access to diverse LLMs.

Files

.env.sample

# Copy this file to `.env` and add your OpenRouter API key.
#   cp .env.sample .env
#
# Get a key at https://openrouter.ai/keys
OPENROUTER_API_KEY=

# LiteLLM model id: openrouter/<vendor>/<model>. Swap the vendor/model freely.
# The default is a `:free` model — runs without purchased credits (rate-limited).
# Paid models (need credits): openrouter/anthropic/claude-opus-4-8, openrouter/openai/gpt-4o
MODEL=openrouter/openai/gpt-oss-120b:free

app.py

"""OpenRouter: one gateway to many models, called through LiteLLM.

Sends one prompt to a model on OpenRouter. The `openrouter/<vendor>/<model>`
prefix routes through OpenRouter with your OPENROUTER_API_KEY, so you can reach
Claude, GPT, Gemini, Llama, and more with a single key by changing MODEL.

Run:
    docker build -t aas-openrouter .
    docker run --rm --env-file .env aas-openrouter "Summarize OpenRouter in one line."
"""

import os
import sys

from dotenv import load_dotenv
from litellm import completion

# Load .env when running locally; in Docker the config comes from --env-file.
load_dotenv()

MODEL = os.environ.get("MODEL", "openrouter/anthropic/claude-opus-4-8")


def main() -> None:
    prompt = " ".join(sys.argv[1:]) or "Summarize what OpenRouter does in one sentence."

    resp = completion(model=MODEL, messages=[{"role": "user", "content": prompt}])
    print(resp.choices[0].message.content)


if __name__ == "__main__":
    main()

Dockerfile

# Minimal image for the OpenRouter sample.
FROM python:3.13-slim

ENV PYTHONUNBUFFERED=1

WORKDIR /app

COPY requirements.txt .
RUN pip install --no-cache-dir -r requirements.txt

COPY app.py .

# Provide your config at run time (.env is NOT baked into the image):
#   docker run --rm --env-file .env aas-openrouter "<prompt>"
ENTRYPOINT ["python", "app.py"]

Related concepts

Related tools

Overview

When to use it

Pricing

When to use

How it flows

When to use

How it flows

Configure

Run with Docker

Run with Docker (in a devcontainer with DooD)

Run locally

Example run