Python client
Create and inspect environments, capture observations, send input, move files, checkpoint work, and manage lifecycle.
Agent connections
Bring the model and framework that fit the work.
Adept AI is being designed as environment infrastructure, not a required agent framework. Initial connection surfaces are proposed and will be verified during early access.
Discuss your use caseProposed interfaces
A small set of explicit controls.
REST API
Use any server-side HTTP client with owner-scoped authentication, idempotent creation, and explicit operation state.
Model adapters
Map the same controls into the agent or model tool interface you already use; keep provider credentials outside the desktop.
PROPOSED API SHAPEIllustrative · no live request
from adept import Client
client = Client()
env = client.environments.create("kicad")
env.wait_until_ready()
frame = env.screen.observe()
env.input.click(x=640, y=412)
env.checkpoints.create()
env.stop()Category breadth was verified against Isle’s public Python and REST documentation, which describes model-neutral HTTP controls and a specific Astra example (source). Adept AI has not released this package or API.
Early access
Plan with us before launch.
Tell us which applications, workloads, and agent stack you need. Early access is selected and subject to technical fit and capacity.