Introduction
Kling O3 Text to video turns a written prompt into a 3-to-15-second clip with no image or reference involved. Choose a length in whole seconds, a resolution from 720p to 4K, and whether to generate native audio alongside the picture. The prompt field takes up to 2,500 characters, enough to lay out the subject, the action, the camera and, with sound on, the dialogue.
Choose a Kling O3 mode by starting material
| What you have at the start | Mode |
|---|---|
| a written scene and nothing else | Kling O3 Text to video |
| an approved opening frame | Kling O3 Image to video |
| up to 7 reference photos | Kling O3 Reference to video |
| a video you already have | Kling O3 Video edit |
Key features & highlights
Multi-shot cuts on its own
Set shot_type to multi and describe more than one shot in the prompt - a wide shot, a close-up, a low angle - and Kling O3 cuts between them in one clip, keeping the setting and the people consistent from one shot to the next. Our rain-soaked tango example is a single prompt describing three shots in order, and the finished clip cut to each one on cue.
Native audio in the same pass
Turn on generate_audio and dialogue, effects and ambience are generated with the picture rather than added afterward. Written dialogue in quotes comes back timed to the action: both lines in our jazz-club example matched the prompt word for word.
720p through 4K, one prompt
All three resolutions render from the same prompt, so the usual pattern is to draft at 720p until the shot and the dialogue work, then re-run the keeper at 1080p or 4K. Resolution and audio both drive the price.
Request parameters
| Parameter | Required | Description |
|---|---|---|
prompt | Yes | Describe the subject, the action, the camera and, with sound on, the dialogue and sounds. For a multi-shot prompt, number the shots in order. at least 1 characters ยท at most 2500 characters |
aspect_ratio | No | 16:9, 9:16 or 1:1. Pick it before writing the shot so the prompt places people and framing for that shape. |
resolution | No | 720p, 1080p or 4K. Higher tiers cost more per second. |
generate_audio | No | Adds a generated soundtrack - dialogue in quotes, effects, ambience - in the same pass. Default: |
duration_seconds | No | Whole seconds, 3 to 15. Drives the price along with resolution and audio. |
shot_type | No | single asks for one continuous take (not a hard guarantee - add "one continuous shot, no cuts" in the prompt if a cut shows up anyway). |
multi_prompt | No | The shot list for |
Pricing
Billed per second of output. Rate depends on resolution and whether native audio is on.
| Resolution | Generate audio: No | Generate audio: Yes |
|---|---|---|
| $0.0672 / second | $0.0896 / second |
| $0.0896 / second | $0.112 / second |
| $0.336 / second | $0.336 / second |
Use cases
Open a scene with no image on hand
Text-to-video is the only Kling O3 mode that needs nothing but words. Use it when the shot exists only in your head, or when you need several takes fast to see which framing works before you commit an opening frame to image-to-video.
Cover a scene in multiple angles from one request
Instead of writing four separate prompts for four angles, describe the sequence once and set shot_type to multi. It is the fastest way to get a wide shot, a close-up and a reaction shot out of a single job.
Prototype dialogue timing before a longer edit
Because audio renders with the picture, a short text-to-video draft is a fast way to check whether a line of dialogue reads naturally against the action, before committing to a longer or higher-resolution run.
Prompt guide & pro tips
Number your shots when you want a cut
For a multi-shot prompt, lay the shots out in the order you want them and describe the camera for each one - wide, close-up, low angle - the way our tango example does. Kling O3 uses that structure to decide where to cut.
Name the camera position, not just the mood
"Seen from the end of the keyboard so both faces are clear" is a real instruction the model can follow; "two men on a piano bench" leaves the angle to chance, and a chance angle can hide one person behind the other. Say where the camera is relative to your subjects whenever more than one has to stay visible.
Ask for a locked shot explicitly
single is a preference, not a guarantee - Kling O3 can still cut on its own. If a scene depends on one unbroken take, add a line such as "one continuous shot, no cuts, the light staying the same throughout" to the end of the prompt.
API quick start
# Generate the idempotency key once; reuse it when retrying after a lost response.
export SPICY_IDEMPOTENCY_KEY="${SPICY_IDEMPOTENCY_KEY:-$(uuidgen)}"
curl -X POST https://api.spicyapi.ai/api/v1/jobs/createTask \
-H "Authorization: Bearer $SPICY_API_KEY" \
-H "Content-Type: application/json" \
-H "Idempotency-Key: $SPICY_IDEMPOTENCY_KEY" \
-d '{
"model": "kling/o3/text-to-video",
"input": {
"prompt": "A slow aerial shot drifting over a misty pine forest at sunrise",
"shot_type": "single",
"resolution": "720p",
"aspect_ratio": "16:9",
"generate_audio": false,
"duration_seconds": 5
}
}'
# Copy data.taskId from the response into TASK_ID, then poll the status.
TASK_ID="TASK_ID_FROM_RESPONSE"
curl "https://api.spicyapi.ai/api/v1/jobs/recordInfo?taskId=$TASK_ID" \
-H "Authorization: Bearer $SPICY_API_KEY"
See the API reference for authentication, the complete schema, polling, and error handling.
