Intake & drafting

You don't start the PRD from a blank page. The Architect agent runs a short intake interview, then streams a complete first draft into the sections live.

Intake

When you open a new project's PRD, the Architect agent asks a short series of intake questions in the chat panel, enough to understand what you're building and for whom. Your answers are saved as you go.

You can answer with a selection or free text. When intake is complete (or you skip ahead), drafting triggers automatically.

Live drafting

Once intake is done, Modus generates the spec section by section and streams each one in as it's written:

1. Problem: the pain point, who it affects, why now
2. Users: one to three personas
3. Scope: what v1 includes and excludes
4. Architecture: the five canonical rows
5. Pages: the screens and their contents
6. User journeys: persona-by-persona paths through the pages

As each section arrives, its tab fills in and a streaming pulse animates so you can watch the spec take shape. The editor auto-jumps to whichever section is currently being written; pick a tab manually and your pick holds until the next section starts streaming.

Confidence

Each section carries an AI confidence level (low, medium, or high) shown as a small badge. It's a hint about where the spec is solid and where you should look harder, not a hard gate.

Redrafting

Not happy with the first draft? Ask the Architect in chat — it can rewrite any section, or the whole spec, over the current inputs.