Pick a tool and ship your first shot

Why this matters

Reading about prompts only gets you so far. At some point you have to pick one video tool, learn its quirks, and generate a clip you can show someone. The tools change monthly — what matters is that you pick one today, ship something, and build the habit. Runway and Kling are the current leaders; Luma, Pika, and Wan are catching up. All of them accept the same structured prompt you've been practicing.

Demo

Shipping a first clip matters more than optimizing it, because iteration feedback only arrives after you generate — reading about prompts can't replace watching a model respond to your specific language. The workflow below walks through tool signup, a copy-pasteable structured prompt, and two comparison generations that reveal model variance without spending extra credits. Thirty minutes from zero to a saved clip is an achievable first session.

1. Sign up for one tool (Runway free tier → 125 credits; Kling free daily credits; Luma's free plan works).
2. Paste this prompt verbatim:
   Subject: a single candle flame in a dark room
   Setting: an empty wooden table, dust motes in the air
   Style: 16mm film, slight halation, warm
   Camera: extreme close-up, static, locked-off
   Lighting: the candle is the only light source
   Motion: the flame flickers and dances for 4 seconds
3. Generate. Wait. Don't tweak yet.
4. Generate a SECOND clip with the SAME prompt. Compare — you'll see the model's variance.
5. Now tweak one line (e.g. "Camera: slow dolly in over 4 seconds"). Regenerate.
6. Download your favorite. That's your first shot.

Try it yourself

1. Sign up for a free tier on any one tool. Don't research all of them — pick in under 5 minutes.
2. Run the candle prompt verbatim. Twice. Compare the variance.
3. Change one line and regenerate. Learn the tool's iteration loop (how long a render takes, what it costs in credits).
4. Ship a clip to one other person — a friend, a Discord, a group chat. You need the feedback loop from day one.
5. Read the tool's prompt guide — each tool has quirks (Kling likes denser prompts, Runway likes negative prompts). One guide, 20 minutes.

Prompt your AI

Use these three in order. Each builds on the one before.

1. Basics & terminology

Compare Runway, Kling, Luma, and Pika as video generation tools — strengths, weaknesses, free-tier shape. Which should I start with if I want cinematic realism? Which if I want anime?

2. Why it works (the mechanism)

What's the difference between a 'text-to-video' generation and 'image-to-video' (where I upload a still first)? When should I reach for each, and why does image-to-video usually give more control?

3. Advanced — application & what's next

Help me turn a 5-second clip I like into a consistent 3-shot sequence — same character, same lighting, three different camera angles. Which tool handles this best right now, and what are the workarounds when it drifts?

References

• Runway — prompt guide
• Kling — text-to-video prompting tips