Outrun API examples

One OpenAI-compatible API key for 96 AI models: GPT, Claude, Gemini, Grok, DeepSeek, GLM, Kimi, Qwen, MiniMax, uncensored and abliterated LLMs, plus image, video and speech models.

Outrun is a prepaid AI API. You top up one USD balance, and every request is billed per use at a published per-token (or per-image, per-second) rate. This repo has copy-paste examples and ready config files for the tools developers already use.

Base URL:  https://api.outrun.to/v1
Auth:      Authorization: Bearer $OUTRUN_API_KEY

• Works with the official OpenAI SDKs (Chat Completions and Responses), the Anthropic Messages API for Claude model IDs, and the Gemini native format.
• The model ID you send is the model that answers.
• Prices: outrun.to/model/ (every model, its billed price, context and setup values).
• Docs: outrun.to/docs/

Quick start

1. Get a key: sign in at outrun.to or open the Telegram bot @outrunAPI_Bot.

2. Put it in an environment variable (never in code or a commit):

   export OUTRUN_API_KEY="your-key"

3. Send a request:

   curl https://api.outrun.to/v1/chat/completions \
     -H "Authorization: Bearer $OUTRUN_API_KEY" \
     -H "Content-Type: application/json" \
     -d '{"model": "gpt-5.6-sol", "messages": [{"role": "user", "content": "Say hello in five words."}]}'

Python (OpenAI SDK)

import os
from openai import OpenAI

client = OpenAI(base_url="https://api.outrun.to/v1", api_key=os.environ["OUTRUN_API_KEY"], max_retries=0)

resp = client.chat.completions.create(
    model="claude-sonnet-5",  # or gpt-5.6-sol, gemini-3.8-flash, grok-4.7, deepseek-v4-pro, glm-5.3-abliterated ...
    messages=[{"role": "user", "content": "Say hello in five words."}],
)
print(resp.choices[0].message.content)

Node (OpenAI SDK)

import OpenAI from "openai";

const client = new OpenAI({ baseURL: "https://api.outrun.to/v1", apiKey: process.env.OUTRUN_API_KEY, maxRetries: 0 });

const resp = await client.chat.completions.create({
  model: "gemini-3.8-flash",
  messages: [{ role: "user", content: "Say hello in five words." }],
});
console.log(resp.choices[0].message.content);

Change only model to switch between GPT, Claude, Gemini, Grok, DeepSeek, GLM, Kimi, Qwen, MiniMax and the uncensored models. All IDs are in CATALOG.md.

The examples set max_retries=0: every request is billed, so if a request is interrupted, check Activity in your dashboard before you send it again.

Examples

Task	curl	Python	Node
Chat Completions	chat.sh	chat.py	chat.mjs
Streaming	stream.sh	stream.py	stream.mjs
Responses API (GPT)	responses.sh
Anthropic Messages (Claude)	anthropic-messages.sh	anthropic_messages.py
Image generation	image.sh	image.py	image.mjs
List models	models.sh
LiteLLM SDK + proxy		sdk_example.py, config.yaml

Video, text-to-speech, speech-to-text and Gemini native examples are in the API reference.

Coding agents and IDE tools

Ready config files, copied from the setup guides in the Outrun docs:

Tool	Config	Protocol	Guide
Claude Code	configs/claude-code	Anthropic Messages	setup/claude
Codex CLI	configs/codex	OpenAI Responses	setup/codex
OpenCode	configs/opencode	OpenAI Chat Completions	setup/opencode
Cline	configs/cline	OpenAI Chat Completions	setup/cline
Roo Code	configs/roo-code	OpenAI Chat Completions	setup/uncensored
Goose	configs/goose	OpenAI Chat Completions	API reference
Continue	configs/continue	OpenAI Chat Completions	API reference

The docs also cover Gemini CLI, Grok Build CLI, the Codex desktop app and Zo Computer.

Models by family

Family	Example model IDs	Page
OpenAI GPT	gpt-6.1-sol, gpt-6-astra, gpt-5.6-sol	GPT API
Anthropic Claude	claude-opus-5-5, claude-sonnet-5, claude-haiku-5-5	Claude API
Google Gemini	gemini-3.8-flash, gemini-3.1-pro-preview	Gemini API
xAI Grok	grok-4.7	Grok API
DeepSeek	deepseek-v4-pro, deepseek-v4.1-flash	DeepSeek API
Zhipu GLM	glm-5.3, glm-5.3-flash	GLM API
Moonshot Kimi	kimi-k3.1-pro, kimi-k3	Kimi API
Alibaba Qwen	qwen3.8-max, qwen3.8-flash, qwen3.7-plus	Qwen API
MiniMax	MiniMax-M3	MiniMax API
Uncensored and abliterated	glm-5.3-abliterated, glm-5.3-uncensored, deepseek-v4-pro-uncensored	Uncensored AI API
Image	gemini-3-pro-image, gpt-image-2, grok-imagine-image	Image setup
Video and speech	veo-3.1, seedance-2.5-720p, gemini-3.1-tts	All models

Model IDs are case-sensitive. The full, current list is in CATALOG.md, regenerated from the public feed https://api.outrun.to/pricing by scripts/update_catalog.py. Prices are kept out of this repo so they can't go stale; the model pages always show the billed rate.

FAQ

Is Outrun OpenAI-compatible? Yes. Set the base URL to https://api.outrun.to/v1 in any OpenAI SDK or tool and use an Outrun key. /v1/chat/completions, /v1/responses, /v1/models, /v1/images/generations, /v1/audio/speech and /v1/videos are supported.

Can I use Claude models with the Anthropic SDK or Claude Code? Yes. Claude model IDs work on /v1/messages with the x-api-key header. In the Anthropic SDK and Claude Code, the base URL is the bare host https://api.outrun.to.

How does billing work? Prepaid. You add credit to one USD balance, and each request is charged at the model's published rate. Eligible new accounts get $0.10 in free API credit to try it. See billing.

Does Outrun store my prompts? Outrun does not store your prompts, images or outputs unless you choose to save them.

Where is a machine-readable model list? https://outrun.to/llms.txt for LLMs and agents, and https://api.outrun.to/pricing as JSON.

Links

• Website: outrun.to
• All models and prices: outrun.to/model/
• Docs: outrun.to/docs/
• llms.txt: outrun.to/llms.txt
• Telegram bot: @outrunAPI_Bot

Found a broken example? Open an issue. Please never paste your API key into an issue.

License

MIT