Content Studio — Reasoning Modal

A step-by-step log of what the AI did during generation — queries fetched, sources considered, personas applied, tokens spent.

What is it?

The Reasoning button in the right-panel action bar opens a modal showing the full generation trace: the analysis step, the personas applied, the queries and pages fetched, the tokens consumed at each step, and the final cost. For Manus drafts it also includes browsing activity and extracted sources. Think of it as an audit trail for the AI's work.

When should I use it?

You want to explain to a client how the content was made.
You're debugging an unexpected result — why did the article skip a key topic?
You want to track token spend per draft for cost analysis.
You're verifying persona usage — did the AI actually apply the persona you selected?

How to use it

Open any generated draft in Content Studio.
Click the Reasoning button in the right-panel action bar (indigo info icon).
Scroll through the step list. Each step shows the action, inputs used, and (for AI steps) input/output tokens.
Close with × or click outside the modal.

Tips and best practices

Check the "Peec.ai analysis" step to see what content gaps were detected before generation.
Token usage rows help spot inefficient prompts — if one step burns 5x more tokens than expected, something's off.
The reasoning log is stored with the draft permanently — you can revisit it months later.

Gotchas and limits

Reasoning is captured only during the generation run. Refinement rounds (Manus chat) don't append to this log.
Very old drafts (before reasoning tracking was added) show a "Not available" message.
Token counts are per-AI-call and don't include the tiny UI-triggered calls (like persona recommendations).
The log is not editable — it's a pure audit artifact.