Features · Scale · Autopilot

Let it run for hours, within a budget.

Name: SHIPIT Agent
Author: SHIPIT

Autopilot turns an Agent or GoalAgent into a long-running autonomous runner: goal-driven, budget-gated, checkpointed every iteration, and streaming as it goes. It runs until every criterion is met or a budget trips.

Autopilot Scheduler daemon

autopilot.pylive

from shipit_agent import Goalfrom shipit_agent.autopilot import Autopilot, BudgetPolicyfrom shipit_agent.llms import build_llm_from_envautopilot = Autopilot(    llm=build_llm_from_env(),    goal=Goal(

$ pythonexample output

running…

How it works

Goal

and success criteria

Iterate

checkpoint every step

Budget

time, calls, tokens, dollars

Result

completed, partial or halted

See it run

The code is a checked example from the docs; the output is an example of what it prints.

autopilot.py

from shipit_agent import Goal
from shipit_agent.autopilot import Autopilot, BudgetPolicy
from shipit_agent.llms import build_llm_from_env

autopilot = Autopilot(
    llm=build_llm_from_env(),
    goal=Goal(
        objective="Summarize three Python dict gotchas with runnable snippets.",
        success_criteria=["3 gotchas explained", "each has a snippet", "each has a tip"],
    ),
    budget=BudgetPolicy(max_seconds=600, max_tool_calls=20, max_dollars=2.0),
)

result = autopilot.run(run_id="py-dict-gotchas")
print(result.status, result.iterations)

$ pythonexample output

running…

Go deeper

4 more ways to use it

Each example is from the docs and checked against the library.

Quick start

Read in the docs

example.py

from shipit_agent import Goal
from shipit_agent.autopilot import Autopilot, BudgetPolicy
from shipit_agent.llms import build_llm_from_env

llm = build_llm_from_env("bedrock")     # defaults to Llama 4 Scout

autopilot = Autopilot(
    llm=llm,
    goal=Goal(
        objective="Summarize three Python dict gotchas with runnable snippets.",
        success_criteria=[
            "At least 3 gotchas explained in prose",
            "Each gotcha has a Python snippet that reproduces it",
            "A one-line avoidance tip is shown for each",
        ],
    ),
    budget=BudgetPolicy(
        max_seconds=600,                 # 10 min cap
        max_tool_calls=20,
        max_iterations=8,
        max_tokens=300_000,
        max_dollars=2.0,
    ),
)

result = autopilot.run(run_id="py-dict-gotchas")
print(result.status)                    # "completed" | "partial" | "halted" | "failed"
print(result.iterations)
print(result.output[:500])

What you get

Budget policy
Caps on time, tool calls, iterations, tokens and dollars.
Crash-safe
An atomic checkpoint every iteration; resume after a crash.
Critic
An optional critic reviews each iteration.
Scheduler
Run autopilots on a schedule with the daemon.

More in Scale

Next: Tracing

from shipit_agent import Goalfrom shipit_agent.autopilot import Autopilot, BudgetPolicyfrom shipit_agent.llms import build_llm_from_envautopilot = Autopilot( llm=build_llm_from_env(), goal=Goal(

See it run

The code is a checked example from the docs; the output is an example of what it prints.

autopilot.py

from shipit_agent import Goal
from shipit_agent.autopilot import Autopilot, BudgetPolicy
from shipit_agent.llms import build_llm_from_env

autopilot = Autopilot(
    llm=build_llm_from_env(),
    goal=Goal(
        objective="Summarize three Python dict gotchas with runnable snippets.",
        success_criteria=["3 gotchas explained", "each has a snippet", "each has a tip"],
    ),
    budget=BudgetPolicy(max_seconds=600, max_tool_calls=20, max_dollars=2.0),
)

result = autopilot.run(run_id="py-dict-gotchas")
print(result.status, result.iterations)

$ pythonexample output

running…

Go deeper

4 more ways to use it

Each example is from the docs and checked against the library.

Quick start

Read in the docs

example.py

from shipit_agent import Goal
from shipit_agent.autopilot import Autopilot, BudgetPolicy
from shipit_agent.llms import build_llm_from_env

llm = build_llm_from_env("bedrock")     # defaults to Llama 4 Scout

autopilot = Autopilot(
    llm=llm,
    goal=Goal(
        objective="Summarize three Python dict gotchas with runnable snippets.",
        success_criteria=[
            "At least 3 gotchas explained in prose",
            "Each gotcha has a Python snippet that reproduces it",
            "A one-line avoidance tip is shown for each",
        ],
    ),
    budget=BudgetPolicy(
        max_seconds=600,                 # 10 min cap
        max_tool_calls=20,
        max_iterations=8,
        max_tokens=300_000,
        max_dollars=2.0,
    ),
)

result = autopilot.run(run_id="py-dict-gotchas")
print(result.status)                    # "completed" | "partial" | "halted" | "failed"
print(result.iterations)
print(result.output[:500])