- A sandbox (the machine the agent runs on)
- A model (the LLM to use)
- A prompt (instructions for the agent)
Create a run
Get a run
Get the result
Oncerun.status is "complete", the final response is available:
from trelent_agents import Client
client = Client()
sandbox = "translator:latest"
run = client.runs.create(
sandbox=sandbox,
model="claude-sonnet-4-5",
prompt="Translate 'Hello world' to Spanish.",
)
from trelent_agents import Client
client = Client()
run = client.runs.get("run_1234567890")
run.status is "complete", the final response is available:
print(run.result)
# => "Hola, mundo."