Features · Build · Streaming

Watch the agent work, event by event.

Name: SHIPIT Agent
Author: SHIPIT

agent.stream() runs the same loop as run() and yields an event for each step as it happens: the model's responses, every tool call and its output, the answer as it is written, and a summary at the end.

Streaming Timeline

stream.pylive

from shipit_agent import Agentfrom shipit_agent.llms import build_llm_from_envagent = Agent.with_builtins(llm=build_llm_from_env(), project_root=".")for event in agent.stream("Run the tests and summarise what fails."):    print(f"{event.type:22} {event.message}")

$ pythonexample output

running…

How it works

agent.stream()

same loop as run()

Events

tool_called, text_delta…

Your UI

terminal, web, notebook

run_completed

the answer and its cost

See it run

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

stream.py

from shipit_agent import Agent
from shipit_agent.llms import build_llm_from_env

agent = Agent.with_builtins(llm=build_llm_from_env(), project_root=".")

for event in agent.stream("Run the tests and summarise what fails."):
    print(f"{event.type:22} {event.message}")

$ pythonexample output

running…

Go deeper

6 more ways to use it

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

An event

Read in the docs

example.py

event.type       # "tool_called"
event.message    # "Tool called: bash"
event.payload    # {"tool": "bash", "call_id": "...", "arguments": {"command": "pytest -q"}}
event.timestamp
event.to_dict()  # for JSON, SSE or WebSocket

What you get

In a terminal
shipit chat renders the same events as a compact transcript.
Over the web
agent.stream_sse() yields Server-Sent Events ready for a response.
From async code
agent.astream() is the async counterpart.
Tool output as it arrives
Long commands stream their output, line by line.

Agent parameters

Parameter	Default	What it does
progress_summaries	False	Emit public `agent_decision` / `agent_observation` progress events. Decisions use only prose the primary model already produced; observations use factual tool metadata.
stream_queue_maxsize	256	How many events `stream()` buffers before the run waits for the reader.
cancel_on_stream_close	True	Cancel the run when the reader stops iterating `stream()`.
stream_join_timeout	2.0	Seconds to wait for the run's thread when a stream closes.

More in Build

Next: Async SDK

from shipit_agent import Agentfrom shipit_agent.llms import build_llm_from_envagent = Agent.with_builtins(llm=build_llm_from_env(), project_root=".")for event in agent.stream("Run the tests and summarise what fails."): print(f"{event.type:22} {event.message}")

See it run

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

stream.py

from shipit_agent import Agent
from shipit_agent.llms import build_llm_from_env

agent = Agent.with_builtins(llm=build_llm_from_env(), project_root=".")

for event in agent.stream("Run the tests and summarise what fails."):
    print(f"{event.type:22} {event.message}")

$ pythonexample output

running…

Go deeper

6 more ways to use it

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

An event

Read in the docs

example.py

event.type       # "tool_called"
event.message    # "Tool called: bash"
event.payload    # {"tool": "bash", "call_id": "...", "arguments": {"command": "pytest -q"}}
event.timestamp
event.to_dict()  # for JSON, SSE or WebSocket

Agent parameters

Parameter	Default	What it does
progress_summaries	False	Emit public `agent_decision` / `agent_observation` progress events. Decisions use only prose the primary model already produced; observations use factual tool metadata.
stream_queue_maxsize	256	How many events `stream()` buffers before the run waits for the reader.
cancel_on_stream_close	True	Cancel the run when the reader stops iterating `stream()`.
stream_join_timeout	2.0	Seconds to wait for the run's thread when a stream closes.