Question 1

What is an AI creative agent?

Accepted Answer

An AI creative agent is software that takes a creative brief and produces finished media by planning the work, picking the right model per step, generating outputs, evaluating them, and iterating — across video, image, and audio modalities. Unlike a single-model generator, an agent orchestrates many models and chains their outputs into a multi-shot project.

Question 2

How is it different from ChatGPT or Midjourney?

Accepted Answer

ChatGPT generates text; Midjourney generates images; a single AI video model generates one type of clip. A creative agent like Tab runs all of those layers — frontier video models, frontier image models, voice, music — and decides which to use per shot. The output is a finished multi-modal project, not just one media file.

Question 3

What does Tab do as an agent?

Accepted Answer

Plans the shot list from your brief, picks Veo 3.1 / Kling 3.0 Pro / Seedance 2.0 / Hailuo 2.3 per shot, drafts the prompts, generates the video, layers ElevenLabs 3.0 voiceover and Suno 5.5 music, and iterates if you change direction. You can drop into Manual mode to take over any step.

Question 4

Can I control what the agent does?

Accepted Answer

Yes. Agent mode runs autonomously for the brief; Chat mode is conversational; Manual mode hands you the model picker. Switch any time — see /agent-mode, /chat-mode, /manual-mode.

Question 5

Which agent model does Tab use?

Accepted Answer

Claude Sonnet 4.6 by default (balanced reasoning, fast). Upgrade to Claude Opus 4.7 for complex briefs or switch to Gemini 3.1 Pro for multimodal-heavy projects. See /claude-vs-gemini, /opus-vs-sonnet.

Question 6

Where do I see this in action?

Accepted Answer

Open /text-to-video and start with Agent mode. For the comparison-style breakdown, see /ai-agent-vs-ai-generator.

Tab | What Is an AI Creative Agent?