T

Text to Video

Create video from text prompt using your central API workflow

Back to Menu Jobs / History

Generate video directly from a structured text prompt

This page sends a text-to-video request through your central Kling job engine. Fill in the prompt, choose your model, set duration and aspect ratio, enter the exact Replicate version ID, and submit the job.

Workflow
Text → Video
Prompt-driven generation
Run Engine
kling-run.php
Shared prediction handler
Status Tracking
Live + Local
Prediction + saved job JSON
Output
Video
Auto download supported

Text to Video Form

Optional internal label for your own job tracking.
Optional preset loaded from kling_profiles.json.
This is your local model label from kling.txt.
Required. The run engine submits by version ID.
Be specific: subject, motion, framing, camera, light, environment, atmosphere and style.
Optional. Add artifacts or elements you want the model to avoid.
Optional. Leave empty if your selected version does not use it.
Optional. Use a seed for repeatability if supported.
Clear Form

Guidance

How to use this page

Choose a Kling model from your local list, paste the exact version ID, write the prompt, optionally add a negative prompt, and submit. The page sends the job through the shared engine and stores a local job record.

Prompt structure

  • Subject: who or what is in the scene.
  • Action: what the subject is doing.
  • Camera: close-up, wide shot, drone shot, tracking shot, dolly, zoom.
  • Environment: city, sea, battlefield, studio, mountains, night scene.
  • Lighting and mood: dramatic, cinematic, foggy, golden hour, realistic.
  • Style: photorealistic, cinematic, documentary, commercial, futuristic.

Version ID note

This page needs the exact version ID because the central run engine submits predictions by version. Keep using the model dropdown as your local label, but always paste the current version ID before running.

Next result check

After submit, use the generated Job Status page link to monitor the job, refresh prediction state, and review local saved output files.