Find your kind of intelligence.
Choose from your available models. Compare capabilities and token rates, then pick the right fit for the task.
Meet the models ↗Rootex Flashrootex-flash
↗✦ A WORKSPACE FOR YOUR NEXT IDEA ↗Great ideas.
Great ideas.
Meet your
new advantage.
Bring AI into what you’re building. Powerful models, a familiar API, and the freedom to work your way.
No monthly subscription Start with $5
rootexAI API studioOne endpoint. Your entire workflow.
PREVIEWYour agent
rootexAI
Response
POST
/v1/responsesJSON{
"model": "rootex-flash",
"input": "Turn this idea into something real.",
"max_output_tokens": 1024
} Your key. Your controls.View documentation ↗
A budget for every key.Your projects stay in your control.
Familiar by designResponses API
The right model for the workFlexible model catalog
Access with boundariesIndividual key budgets
A clearer view of your workRequest-level analytics
A LITTLE LESS SETUP. A LOT MORE POSSIBILITY.Everything you need.
Everything you need.
Nothing in your way.
From your first experiment to your everyday workflow, keep the useful things close.
Give every project its space.
A separate key for each agent. Set total or daily budgets, adjust the limits, and revoke access whenever you need.
Your keys, your rules ↗Total + daily limitsIndependent budgets per API key
Know where every token goes.
See model activity, tokens, timing, and spending for each key. The details you need, without a spreadsheet.
A clearer picture ↗From request to reportModels · tokens · cost · latency
↗A FAMILIAR STARTING POINTYour tools.
Open the quickstart guide ↗
Your tools.
A new possibility.
Use a Responses-compatible agent or SDK. Add your rootexAI key and base URL, choose a model, and get back to the work that matters.
01
Create an API key with its own budget.
02
Set your base URL and public model ID.
03
Make a request. See it in your analytics.
PythonResponses API
from openai import OpenAI
client = OpenAI(
api_key="YOUR_API_KEY",
base_url="YOUR_SITE_URL/v1"
)
response = client.responses.create(
model="rootex-flash",
input="Let’s build something useful."
)
print(response.output_text) Your app connects to rootexAI. Routing stays on the server.