Control and security

See exactly what the agent did, step by step

The Inspector in Brainwrite is a detailed view of a conversation that shows tool calls, reconnect attempts, engine output and long-running work in more depth than the chat. You open it from the chat's header menu in Advanced mode. Export saves the conversation, its events and the engine's raw messages in one file, in time order, with secrets removed.

Download for Mac Read the docs

Why did the run stall?

Prepare release notes for v2.3 from the merged pull requests.

Your Chief of Staff put a team on it

Nova· Chief of Staff
Dash· Triage bot
Atlas· Rigel, QA and Release Engineer

Tool call: list merged pull requests, 23 results.
Stream dropped; reconnect attempt 1 succeeded.
Engine error: 401, the CLI sign-in has expired.
Export saved with secrets removed.

DoneSign in to the engine's CLI again, then retry: waiting for your yes.

An illustrated example. Your team works on your own files and apps.

A lawyer highlights a contract at her desk while illustrated AI agents Nova and Atlas stand beside her with tabbed papers.

See it in the app

Inspector in Brainwrite.

The Inspector beside a routine run: Run Log, Events and Raw tabs listing user input, three completed tool calls and the recorded response.

How it works

How Inspector works.

01
Turn on Advanced mode
Open Settings → General → Advanced mode. The choice applies to this device and only shows or hides controls.
02
Open the Inspector
Choose Inspector from the chat's header menu.
03
Follow the run
Watch tool calls, engine output and reconnect attempts as the bot works, or review them afterwards.
04
Export if you need to share
Choose Export to save one file with the conversation, its events and the raw engine messages. Review it before sending.

Good to know

The details.

Secrets removed, content kept

Export strips secrets but keeps what was said and done, in time order.

Retries you can see

Brainwrite retries transient streaming failures. If retries end in an authentication or configuration error, fix the engine rather than retrying the task.

Everything stays with its task

Questions, choices, permission requests, tool activity and failures remain attached to the task that produced them.

Made for bug reports

The exported file is the thing to attach when you report a problem, because it shows the engine's own messages, not just the chat.

The edges, plainly

The Inspector is available in Advanced mode only.
Exports include the full conversation and what the agent did. Review the file before you share it.

Full reference in the documentation

Use cases

Jobs this feature does for you.

Works with

Apps it's often used with.

FAQ

Questions about Inspector

How do I see which tools an AI agent called?

Open the Inspector from the chat's header menu. It follows tool calls, reconnect attempts, engine output and long-running work in more detail than the conversation shows. The Inspector is part of Advanced mode, which you turn on in Settings → General → Advanced mode.

How do I debug an AI agent that keeps failing?

Open the Inspector to see the tool calls and raw engine output. Brainwrite already retries transient streaming failures, so if retries end in an authentication or configuration error, fix the engine itself, such as its CLI login or executable path, rather than retrying the task.

Does an exported AI agent conversation include my API keys?

No. Export removes secrets. It does include the conversation, its events and the engine's raw messages in time order, so personal details or file contents that came up in the chat may be in it. Review the file before you share it.

Works well with