5/10 seconds · optional audio · text or reference image
Every example below is an actual Kling 2.6 task run through this workspace. We list the submitted mode, duration, audio setting, charged credits, and the dimensions measured from the saved MP4 instead of promising a fixed resolution.
APIMart UI integration test: a silver toy robot walking slowly across a dark blue studio floor, fixed camera
A small paper sailboat glides across a shallow mirror-like pool at sunrise. Slow cinematic push-in, gentle ripples, soft wind and water ambience, no text, no logos.
The working model route exposes two generation modes. The selected controls and debit shown in the generator are the source of truth for each task.
Write a scene prompt, choose 5 or 10 seconds, select 1:1, 16:9, or 9:16, and optionally enable synchronized audio with the Sound control. Camera direction can be written in the prompt; there is no separate camera-path editor.
Open the video generatorUpload one reference image, describe the motion, choose 5 or 10 seconds, and optionally enable synchronized audio with the Sound control. The reference image anchors the scene; the current UI does not accept an audio file as an input.
Open the video generatorKling 2.6 costs vary by duration and whether audio is enabled. The number beside Generate updates with the selected settings and is the authoritative debit. Failed eligible tasks follow the bounded refund flow.
View credit packsThe usable generator comes first on this page. Complete the task in three transparent steps.
Start with a written scene, or switch to Reference to Video and upload one image you have permission to use. Add subject motion, scene changes, and camera direction in the prompt.
Choose 5 or 10 seconds. Text-to-video also exposes 1:1, 16:9, and 9:16. Turn Sound on only when you want the same task to generate synchronized audio, then review the displayed credits.
Sign in, submit the task, and keep the page open or return to task history. Successful MP4 output can be played and downloaded; the task record shows the actual debit and status.
Use short generated clips as drafts and creative inputs, then review motion, audio, continuity, rights, and factual details before publishing.
Turn a product description or approved product image into a short motion study for a launch, ad, or storyboard.
Test vertical, square, and landscape framing before committing to a channel-specific edit.
Describe a pan, push-in, orbit, handheld feel, or fixed camera in the prompt and review the actual result.
Enable audio for one synchronized output, then verify dialogue, ambience, and timing before reuse.
Animate an approved still while keeping the source as the visual anchor for the generated clip.
Explore blocking, pacing, and composition as an early draft—not as guaranteed final footage.
Verified answers about the current generator, inputs, audio, output files, credits, and site ownership.
No. Kling26.app is an independent third-party workspace. It is not operated by, affiliated with, or endorsed by Kuaishou or Kling AI, and it is not the official Kling website.
The current UI supports text-to-video and reference-image-to-video. Both require a text prompt. Image-to-video accepts one reference image; the current UI does not accept an audio file, existing video, motion path, or keyframe sequence as input.
Kling 2.6 exposes 5- and 10-second durations. Text-to-video offers 1:1, 16:9, and 9:16. Image-to-video uses the reference image to guide framing and does not expose a separate aspect-ratio selector in the current route.
Yes, when Sound is enabled. Audio is generated with the video task and returned in the resulting MP4 when the provider succeeds. It is not a separately downloadable stem, and the site does not offer custom voice training.
You can describe camera behavior—such as a slow push-in, orbit, pan, or fixed camera—in the prompt. The current workspace does not provide a separate motion-path or keyframe editor, so results must be reviewed rather than assumed.
The current Kling 2.6 route does not expose a resolution or export-format control. Successful results are downloadable MP4 files. Output examples on this page show dimensions measured from each saved file; ProRes, frame sequences, and separate audio stems are not offered.
The debit depends on duration and whether Sound is enabled. The credit number beside Generate updates before submission and is the authoritative price. Each real example records the amount actually charged to that task.
Signed-in users can play or download a finished result in the generator and revisit saved tasks in Activity. Provider links may be temporary, so saved site-controlled output is preferred when available.
Choose the working mode, verify the displayed duration, audio setting, and credits, then generate and review the actual MP4 output.
