Keiro API platform
On this page 1 of 9
Keiro eb1 API
Make your first request in minutes
Keiro is OpenAI-compatible. Change three lines in your existing client — the base URL, your key, and the model — and you are running on eb1.
from getpass import getpass
from openai import OpenAI
client = OpenAI(
base_url="https://api.keirolabs.ai/v1",
api_key=getpass("Keiro API key: "),
)
response = client.responses.create(
model="eb1-preview",
input="Write a deployment checklist.",
)
print(response.output_text)
Start
Build
API surfaces
The API reference is the index for all public endpoints.
All completion surfaces support streaming. Tool and image features use the native request shape of each endpoint.
Start building#
- Quickstart goes from key creation to a first response in a few minutes.
- Models compares the four public eb1 model IDs.
- Run your coding agent on eb1: Claude Code, Codex, or Omnigent.
Operate an account#
- Usage and billing explains console totals, request logs, exports, and pending versus finalized records.
- Rate and spend limits documents limit headers, spend caps, and retry timing.
- Errors lists machine-readable codes and stream-error fields.
- Status and support gives the safe incident-reporting path.
Paid-beta boundary#
The console manages organizations, keys, models, usage, billing, limits, and support access. The local keiro gui command is for chat exploration only.
Public docs describe supported customer contracts. They do not expose or promise private routing, provider selection, or operator interfaces.
Trust and policy#
Review Legal and privacy, Data retention, Beta SLA, and Subprocessors before sending a sensitive workload. During the paid beta, supported API request and terminal response payloads are retained by Keiro Labs to operate and improve the service; see Data retention.
Models
Keiro Étude is the first generation of eb1. The next is already in training.