All products
OpenBot API
Automate the loop data workflow with one API surface.
Queue rollouts, inspect failures, and feed results back into curation. Use REST, Python, or agent tool schemas.
- 3
- lines to start
- OpenAPI
- 3.1 spec
- Python
- SDK
# Python SDK
from openbot_sdk import Client
ob = Client()
result = ob.bench.rollout(
policy="openvla-7b",
embodiment="franka_panda",
task="kitchen_handover",
).wait()
print(result.task_success)Quick start
From key to failure signal.
01
Get an API key
Create a key in the dashboard and scope it to current Data or Bench operations.
02
Queue a rollout
POST a policy, embodiment, and task. Runs are async and idempotent.
03
Get metrics
Poll or receive task success, intervention rate, and sim-to-real gap.
Developer experience
Three ways to call. One surface.
REST for curl scripts, Python SDK for notebooks, TypeScript SDK for dashboards. Same endpoints, same types, same async behavior.
rollout.python
from openbot_sdk import Client
ob = Client() # reads OPENBOT_API_KEY
run = ob.bench.rollout(
policy="openvla-7b",
embodiment="franka_panda",
task="kitchen_handover",
rollouts=200,
webhook="https://my.app/openbot",
)
result = run.wait()
print(result.task_success) # 0.73Use cases
Built for agents, CI, and async workflows.
LLM agents
Give agents a policy-evaluation tool before they act.
CI pipelines
Block regressions and send failures back into the data loop.
Async workflows
Queue long evaluations and receive a webhook when done.
Ready to integrate?
Read the reference, install the SDK, or request a live key.