Look like you. Sound like you.

1. 01

   Identity intake

   Upload photos, a description or an approved render and pick the style. Confirm the identity record: face, eyes, hair, skin, proportions, garments, footwear, accessories, palette and the details that make you you. No sensitive traits are inferred from photos.

2. 02

   Canonical anchor

   One full-body render plus a face anchor. You approve it before anything else is generated.

3. 03

   Character card

   Full body on the left third, a large head-and-shoulders bust on the right, on a plain off-white background.

4. 04

   Identity views first

   Turnaround (5), head construction (3) and silhouettes (4), inspected before expressions and poses start.

5. 05

   Expressions, visemes, poses

   Each generated as its own asset with your canonical anchors attached.

6. 06

   Board assembly in code

   Fixed layout, labels and borders composed exactly, not drawn by the image model.

7. 07

   Continuity QA

   Every asset marked pass, fail or unverified across nine areas: identity, anatomy, wardrobe, orientation, style, function, speech, layout and motion. Failed assets are repaired one at a time from the canonical references.

8. 08

   Lock

   The status badge moves from DRAFT to IDENTITY LOCKED once the anchor is approved and the pack passes.

The three sheets

One fixed template, rendered at 3344x1882. Slots that do not apply show N/A with a reason; no slot is deleted.

1. 01

   Consent and path

   Your own voice, or written permission from the speaker. Pick Instant or the video-platform voice.

2. 02

   Equipment and room checklist

   USB or XLR mic, pop filter at two fists, a small soft room, fans and notifications off, headphones.

3. 03

   Record in the browser

   Mono, 48 kHz, peaks at -12 to -6 dB with a live meter that warns at 0 dB, and 5 seconds of room tone first.

4. 04

   Sample script

   A 380-word script covering all 39 American-English sounds, with your name and your own frequent words dropped in.

5. 05

   Clean up

   Cut flubs in the in-app trimmer, then the server filters, trims pauses, levels and limits the sample. Your raw recording is kept.

6. 06

   Check

   A report on length, loudness, peak, clipping, noise floor, long pauses, channels and format, each row PASS, FAIL or WARN.

7. 07

   Create the clone

   Your cleaned sample becomes your voice ID.

8. 08

   Test

   Your clone reads a test paragraph that was not in the sample; you read the same paragraph; you hear both side by side.

9. 09

   Fix

   A seven-row troubleshooting table (echo, hiss, pace, monotone, drifting accent, a mispronounced word, doesn't sound like me), each linked to the step to redo.

Your voice record holds the provider, voice ID, sample file, settings, check report and test clip. That record is all Video and Agents need.

Presenter profile, once per style

2D, 3D or Photoreal, each built from its locked Image pack and your Voice record.

1. 01

   Identity block

   Fixed text pasted unchanged into every presenter prompt, generated from your Image clone's prompt block.

2. 02

   Two keyframes

   Full frame, mid-chest, in a setting that suits the topic; and side, in the left third with an open palm toward an empty right side.

3. 03

   Pilot

   One full-frame line, one presenter-beside-visual line, the hardest motion and one state transition, cut into a short proof reel that you approve.

Video Studio, per episode

Structure: hook, intro with your name, context, one idea per step, takeaway, call to action. On-camera lines run 14 seconds or less.

1. 01

   Collect inputs

   Topic, audience, a one-sentence takeaway, length, presenter style, voice, brand colors, font and logo, call to action, and a source for every factual claim.

2. 02

   Plan (no credits)

   Decisions table, brief, script table, storyboard with the Voice version visual per scene, two short-form scripts and a claim and source map. You approve the script and storyboard.

3. 03

   Visual system

   Chosen by topic: state frames for a process, charts drawn in code for data, screen recordings for software, illustrated scenes for a story, split screens for a comparison, one metaphor object for an abstract idea.

4. 04

   Frames

   Master and state frames generated and reviewed: the subject keeps its outline, change happens only where intended, no stray text.

5. 05

   Batch

   Narration for every scene first, then stills, motion clips and lip-synced presenter clips in parallel. Every job is priced before it is submitted.

6. 06

   Assemble both versions

   One script, two cuts, built scene by scene.

7. 07

   Verify

   Both masters must pass the audio, narration-match and silence checks, and both contact sheets are reviewed. The Voice version shows no presenter frame.

8. 08

   Package

   Two upload encodes, four vertical short clips, captions, two thumbnails, upload copy with chapters per version, and a manifest of every job and its credits.

1. 01

   Brief

   A fill-in project brief: space type, purpose and audience, the spaces and features that must appear, route, duration and a spending limit. Nothing is required; sensible defaults fill the blanks.

2. 02

   References

   Upload photos, videos, floor plans, drawings or renderings. Videos are sampled into stills, each with its own ID.

3. 03

   Plan

   A reference map, a coverage matrix, the runtime and a shot plan with one prompt per shot. Every shot is editable before you approve the plan.

4. 04

   Generate

   Each shot renders once from its stills, with the price shown first and your limit respected. You accept each shot or mark it defective with what is wrong.

5. 05

   Assemble and export

   Accepted shots are joined with cuts or dissolves at a constant frame rate into an H.264 MP4 that starts playing fast.

6. 06

   Validate

   Technical checks on the exported file, a frame-by-frame visual check of every shot and transition, a playback check and a completeness check.

7. 07

   Deliver

   Your video inline with a direct download, plus any coverage gaps stated plainly. Nothing is invented to fill a gap.