Quickstart
If you've used OpenAI, you've used Aigarth.
from aigarth import Aigarth
client = Aigarth(api_key="sk-aigarth-...")
response = client.chat.create(
model="aigarth-reason-1",
messages=[
{"role": "user", "content": "Explain Useful Proof of Staking in 3 sentences."}
],
)
print(response.choices[0].message.content)SDKs for every stack
Idiomatic libraries in the languages your team already uses.
Python
pip install aigarthTypeScript
npm install aigarthGo
go get aigarth.cloud/sdkRust
cargo add aigarthJava
implementation 'cloud.aigarth:sdk'C#
dotnet add package AigarthOpenAI-compatible
Same endpoints, same request shape, same response format. Switch by changing the base URL.
Powerful CLI
Auth, run, deploy, monitor, and govern from your terminal. Scriptable for CI/CD.
Streaming & async
First-class streaming, async/await, and tool use. Built for production applications.
Scoped keys
Per-project keys, environment separation, IP allowlists, and per-key spend caps.
Live playground
Try any model from your browser. Inspect latency, tokens, and cost in real time.
Comprehensive docs
API reference, guides, recipes, examples. Searchable, versioned, complete.
Open source
SDKs, CLI, and core protocol are open source. Inspect, fork, contribute.
Python SDK with sync/async, streaming, and full type hints.
TypeScript SDK. Works in Node, Bun, Deno, and the browser.
Command-line interface for auth, deploy, monitor, and govern.
Core protocol implementation. Reference for network operators.