Features · Scale · Tracing

Every run, recorded for later.

Name: SHIPIT Agent
Author: SHIPIT

Tracing captures every event emitted during a run, from tool calls and retries to the final output, into a persistent store you can reload, inspect and compare later.

Tracing Observability exports

tracing.pylive

from shipit_agent import Agent, FileTraceStoreagent = Agent.with_builtins(    llm=llm,    trace_store=FileTraceStore(root_dir=".shipit_traces"),    trace_id="run-001",)

$ pythonexample output

running…

How it works

Run

every event

Trace store

on disk or yours

Reload

by trace id

Inspect

pretty-print or compare

See it run

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

tracing.py

from shipit_agent import Agent, FileTraceStore

agent = Agent.with_builtins(
    llm=llm,
    trace_store=FileTraceStore(root_dir=".shipit_traces"),
    trace_id="run-001",
)

result = agent.run("Search for Python 3.13 release notes")

$ pythonexample output

running…

Go deeper

6 more ways to use it

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

Quick start

Read in the docs

That's it. Every event is now persisted to `.shipit_traces/run-001.json`.

example.py

from shipit_agent import Agent, FileTraceStore

agent = Agent.with_builtins(
    llm=llm,
    trace_store=FileTraceStore(root_dir=".shipit_traces"),
    trace_id="run-001",
)

result = agent.run("Search for Python 3.13 release notes")

What you get

Trace stores
On disk, in memory, or your own store.
Pretty-print
Read a run as a tree, or as raw JSON.
Compare traces
Diff two runs of the same task.
OpenTelemetry
Export spans to the observability stack you already use.

Agent parameters

Parameter	Default	What it does
trace_store	None	Where run traces are written.
trace_id	None	The trace id for this agent's runs.
observability	'auto'	Report this agent's runs through shipit-watcher — traces, the agent graph, scores, datasets and the local ledger. `True` attaches it, `False` never does, and the default `"auto"` attaches it only when shipit-watcher is installed and configured with credentials.

More in Scale

Next: Any model

See it run

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

tracing.py

from shipit_agent import Agent, FileTraceStore

agent = Agent.with_builtins(
    llm=llm,
    trace_store=FileTraceStore(root_dir=".shipit_traces"),
    trace_id="run-001",
)

result = agent.run("Search for Python 3.13 release notes")

$ pythonexample output

running…

Go deeper

6 more ways to use it

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

Quick start

Read in the docs

That's it. Every event is now persisted to `.shipit_traces/run-001.json`.

example.py

from shipit_agent import Agent, FileTraceStore

agent = Agent.with_builtins(
    llm=llm,
    trace_store=FileTraceStore(root_dir=".shipit_traces"),
    trace_id="run-001",
)

result = agent.run("Search for Python 3.13 release notes")

Agent parameters

Parameter	Default	What it does
trace_store	None	Where run traces are written.
trace_id	None	The trace id for this agent's runs.
observability	'auto'	Report this agent's runs through shipit-watcher — traces, the agent graph, scores, datasets and the local ledger. `True` attaches it, `False` never does, and the default `"auto"` attaches it only when shipit-watcher is installed and configured with credentials.