One API for leading AI models. Route requests by cost, quality or speed—and keep control of your spend.
OpenAI-compatible · No routing fee
Compare the models behind Ninja Router. Set your token usage and see what each costs.
Connect to more models with one integration. Keep your inference costs under control.
One integration for your agents, apps and workflows. Choose the model when you want to. Let us route when you don’t.
What could your team do with 40% less inference spend? More experiments. More runway. More room for the next launch.
Prioritize cost, answer quality or response speed. Use a fixed model whenever you need one.
Open developer consoleCompare monthly costs at your request volume.
Based on 5K input + 1K output tokens per request at current rates.
Keep your OpenAI-compatible integration. Use your NinjaChat key, update the base URL, and set your model to ninja/auto.
Set NINJACHAT_API_KEY in your environment.
01import os
02from openai import OpenAI
03
04client = OpenAI(
05 api_key=os.environ["NINJACHAT_API_KEY"],
06 base_url="https://www.ninjachat.ai/api/v1",
07)
08
09response = client.chat.completions.create(
10 model="ninja/auto",
11 messages=[{"role": "user", "content": "Summarize this: Our launch is Friday."}],
12 extra_body={"routing": {"strategy": "cost"}},
13)
14print(response.choices[0].message.content)