A brief welcome or introduction
Pair a clear portrait with a recorded greeting for an introduction, tutorial opening or community update.
Talking Avatar
Make a talking photo from an image and your audio
Choose a portrait, upload or record up to 15 seconds of audio, and describe the avatar’s visual performance. Generate a new talking clip and review its expression and timing.
History
Your recent lip-sync videos appear here.
Sign in to start
Example: record a short launch greeting
Prepare a portrait and record “Welcome—here is a quick look at our new collection.” Leave room for the final word to finish within 15 seconds.
Upload the portrait and audio, or select Use recording after checking the microphone draft. Wait until both inputs are ready. Add visual guidance such as “A relaxed presenter, a small friendly smile, one gentle nod, steady framing.” Generate only after checking these inputs.
Open the result from history and watch it with sound. Check lip timing, facial drift and movement. Download the take you want to keep. Use the audio cutter to prepare a shorter source, or browse pricing and all applications.
Keep the message short and focused
A product announcement
Use a short voice clip to introduce a launch or feature. Review names, mouth movement and expression before adding it to your edit.
A character performance test
Try a short line to explore a character’s expression and gesture. Review likeness and movement.
Prepare your material or finish the edit
What you provide
One JPG, PNG or WebP image, up to 10 MiB.
One MP3, WAV, AAC, M4A or OGG audio file, up to 10 MiB and 15 seconds, or a short microphone recording.
A required visual prompt of up to 5,000 characters to guide expression, gesture and scene treatment.
History lets you check status, play results, open them separately and download available videos.
Before creating a talking avatar
Why is the prompt required if I already have audio?
Audio supplies the words and delivery. Use the prompt for visible behavior: a calm expression, slight nod or steady camera. It does not create speech, translate or rewrite the recording.
What kind of image works best?
Use a well-lit portrait with a clear face and unobstructed mouth. Extreme angles or tiny faces can animate poorly; inspect teeth, lips and facial details.
What kind of audio should I prepare?
Use a clear, short recording with one speaker and limited background noise. The page accepts readable audio up to 15 seconds; trim a longer take before uploading. Record the wording and pacing you want instead of trying to change them through the visual prompt.
How do I use the microphone?
Sign in, choose Record audio and allow microphone access. Stop within the limit and listen. Discard or record again if needed. Use recording uploads the take; it does not start generation.
Can I dub an existing video or generate a voice here?
No. This page creates a new clip from a still image and supplied audio. It has no existing-video input, voice library, translation or voice-cloning controls. Prepare an audio file first if your starting point is a written script.
What uses credits, and can I retry?
Sign-in is required for uploads and recording. Generation uses cloud processing and account credits; consult current pricing. Refresh checks the current task without generating another clip. A new Generate request can charge again; there is no task Retry button or guaranteed completion time.
Can I choose the model or export settings?
The page uses Kling AI Avatar Pro without a model, resolution, duration or export-format selector. Review and download the video. Check the sound and the end of the last word as well as the face.