← Back to blog · 6 min read · Published 2026-03-08 · Updated 2026-04-20 · By Flik AI
Manual Mode: The Power User's AI Video Workflow
When you want full control — pick the model, write the prompt, tune every parameter yourself.
Manual mode is Flik AI for power users. No agent interpreting your brief, no automatic model routing, no AI-written prompts — you pick the model, write exact prompts, tune the aspect ratio and duration, and generate. For production work where reproducibility and precision matter, this is the right mode.
When Manual mode beats Agent or Chat
- You're producing at scale and need the same prompt to produce consistent output
- You've spent time crafting a prompt you want to reuse across a catalog
- You're a video professional who knows exactly what camera direction, lighting, and style you want
- You need to match an existing visual spec (brand guidelines, campaign style) with zero AI deviation
- You're iterating on a technical-parameter level (aspect ratio, duration, resolution, reference weights)
What you control in Manual mode
Manual mode parameters per model| Parameter | Where it applies | Notes |
|---|
| Model selection | All generations | Pick Veo 3.1, Kling 3.0 Pro, Kling 2.6, Kling o3 Edit, Kling o3 Ref, Seedance 2.0, Seedance 2.0 Fast, Hailuo 2.3, or any image/audio model |
| Prompt text | All generations | Full 6-part formula (subject, context, action, style, camera, lighting) |
| Aspect ratio | Video + image | 16:9, 9:16, 1:1, 4:5, 3:4, 2:3, 21:9, 4:3, 5:4 |
| Duration | Video | Varies per model — up to 15s on Seedance, ~8s on Veo, 60+ on Kling 3.0 Pro stitched |
| References | Multi-ref models | Up to 9 images + 3 videos + 3 audio on Seedance 2.0 |
| Generation count | All | x1, x2, x3, x4 variants per generation |
Prompting in Manual mode
Use the 6-part cinematographer's formula: subject, context, action, style, camera, lighting. Each model has specific tips on top of the formula — see /veo-3-prompts, /kling-prompts, /seedance-prompts, /hailuo-prompts, /seedream-prompts, /nano-banana-prompts, /elevenlabs-prompts, /suno-prompts for model-specific guides. The full prompting framework is at /blog/how-to-prompt-ai-video.
Manual mode production tips
- Draft with cheap-tier models (Kling 2.6, Seedance 2.0 Fast) to find the right prompt, then final-render with premium (Kling 3.0 Pro, Veo 3.1)
- Save winning prompts as project templates for reuse across catalog or campaigns
- Use consistent reference images across shots to lock visual style — especially with Seedance 2.0's 9-image system
- For iteration, change one prompt variable at a time (lighting only, camera only) and compare A/B
- Export all aspect ratios in parallel when the prompt is dialed — don't re-prompt per platform
For the Manual mode marketing page, see /manual-mode. For autonomous end-to-end, see /agent-mode.
Tags: manual mode power user how-to workflow
Frequently asked questions
Which models are available in Manual mode?
Every supported model: Veo 3.1, Kling 3.0 Pro, Kling 2.6, Kling o3 Edit, Kling o3 Ref, Seedance 2.0, Seedance 2.0 Fast, Hailuo 2.3, Seedream 4.5, Nano Banana Pro, Nano Banana 2, ElevenLabs 3.0, Suno 5.0, and Mirelo 1.5. Agent/chat models (Claude, Gemini) aren't used in Manual mode — they're for Agent and Chat modes.
Is Manual mode harder than Agent mode?
Yes — Manual mode requires writing prompts and picking models yourself. But it gives reproducibility and precision that Agent mode's autonomy can't match. Most power users switch between modes depending on the task.
Can I save prompts for reuse in Manual mode?
Yes. Save any project as a template and reuse its prompt + references + model settings. Pair this with Seedance 2.0's 9-image reference system to lock brand style across an entire catalog of generations.
Manual mode vs Agent mode — which is cheaper?
Per-generation cost is the same across modes. Manual mode often ends up cheaper in total because you iterate more efficiently (no agent-token overhead on refinement chat), especially when you're comfortable writing prompts.
Related posts
- Agent Mode: The Complete Guide to Flik AI's Autonomous Workflow — Agent mode turns a plain-English brief into a finished multi-model AI video. Here's the complete guide — how it plans shots, picks models, writes prompts, and iterates.
- Chat Mode: The Iterative AI Video Workflow — Chat mode is Flik AI's conversational creative loop — iterate turn-by-turn with the agent, steering between generations without rewriting prompts.
- How to Prompt AI Video: The Complete 2026 Framework — The six-part prompt structure used inside Flik AI — subject, context, action, style, camera, lighting — with real examples across Veo 3.1, Kling 3.0 Pro, Seedance 2.0, and Hailuo 2.3 plus model-specific prompting tips.
Try Flik AI · More posts · FAQ · Pricing
Home · AI Video Generator · Text to Video · Image to Video · Veo 3.1 · Seedance 2.0 · Kling 3.0 Pro · Seedream 4.5 · ElevenLabs 3.0 · Suno 5.0 · Pricing
© 2026 Flik AI. All rights reserved.