Text-to-Video Generation

Text-to-video models create video from a text description alone — no source clip or image is required. This is useful for generating establishing shots, abstract visuals, motion graphics backgrounds, or any footage that does not exist yet.

Step-by-Step

1. Open the Panel

In Premiere Pro, go to Window > Extensions > modelBridge.ai.

2. Search for a Text-to-Video Model

In the model selector, search for a text-to-video model. Some recommended options:

Model	Strengths	Typical Cost
Kling v3 Pro	High quality, consistent motion	~$1.00–2.00 per clip
Wan 2.6	Good balance of quality and speed	~$0.20–0.50 per clip
LTX 2.3	Fast generation, lower cost	~$0.05–0.15 per clip
Veo 3	High fidelity, longer clips	~$1.50–3.00 per clip

Use the filter chips to show only text-to-video models.

3. Write Your Prompt

Enter a text prompt that describes the video you want. Text-to-video models respond well to specific, descriptive prompts.

Good prompt example:

“Aerial drone shot flying over a turquoise ocean at sunset, gentle waves, golden hour lighting, cinematic color grading, 4K quality”

Weak prompt example:

“Ocean sunset”

Tips for effective prompts:

Describe camera motion — “slow pan left”, “dolly forward”, “static wide shot”
Specify lighting — “golden hour”, “neon-lit”, “overcast soft light”
Include style cues — “cinematic”, “documentary”, “hand-held”
Mention mood and atmosphere — “peaceful”, “tense”, “dreamlike”
Be specific about subjects — “a woman in a red coat walking through rain” rather than “a person walking”

4. Set Duration and Resolution

Adjust the parameters in the form below the prompt:

Duration — how long the generated video should be (typically 2–10 seconds depending on the model)
Resolution — output resolution (720p, 1080p, etc.)
Aspect ratio — match your project settings (16:9, 9:16, 1:1)

Higher duration and resolution increase both generation time and cost.

5. Check the Cost Estimate

Look at the cost badge near the Generate button. It updates live as you change parameters. If the cost seems high, try reducing duration or resolution first.

6. Generate

Click Generate. The button shows progress stages:

Submitting your request
Queued (waiting for GPU)
Generating (AI is producing frames)
Downloading the result
Importing to Premiere Pro

Generation time varies by model — from 15 seconds to several minutes. If it takes too long, the generation moves to the background automatically so you can continue editing.

7. Use the Result

The generated video appears in your Project panel automatically. From there you can:

Drag it onto the timeline
Preview it in the Source Monitor
Generate additional variations with different prompts

No Source Clip Required

Text-to-video models do not need a clip selected on the timeline. The media card will show that no selection is needed, and the Generate button will be active as long as you have entered a prompt.

If you want to generate video from an existing image instead, see the Image-to-Video workflow or use an image-to-video model.

Common Issues

Generation takes too long

Some models queue during peak hours. If the status stays on “Queued” for more than a few minutes, try again later or switch to a faster model like LTX 2.3.

Result does not match the prompt

AI models interpret prompts differently. Try rephrasing, adding more detail, or switching to a different model. Each model has its own strengths and visual style.

Cost higher than expected

Enable audio, high resolution, or long duration can multiply costs. Check the cost badge before generating and reduce parameters for test runs.