Skip to main content

Kling Avatar 2.0 talking avatar turns one portrait and an audio file into a clip of that person speaking or singing it. Render at 720p or 1080p; the clip length follows your audio, up to about 5 minutes.

From $0.056/s

Workbench

Input

Talking avatar
183/2500
Optional direction for the performance: mood, gestures, where the eyes go. Separate from what the audio says. at most 2500 characters
1/1

Drop files or media links, paste, or browse · 1 file · Max 10 MB · JPG · PNG

  • 1
1 / 1 files
Portrait to animate. JPEG or PNG only. A clear, front-facing, well-lit face gives the most reliable lip sync.
1/1
1 / 1 files
Speech or song the avatar performs. The clip is exactly as long as this audio - there is no separate duration field.
affects price
720p or 1080p. 1080p costs more per second of audio.
Safety checkerOff

SpicyAPI does not run a safety checker on this model. Your prompt and media go to the model as written, and the model itself decides what comes back.

Checking your account before enabling submission and pricing.

Output

Ready for a new run
Example result

Missing the model you need?

TALKING AVATAR VIDEO

Kling Avatar 2.0 API at a glance

Send a portrait and a line of speech or a song, and get back a clip of that person saying or singing it - lip sync, expression and length all follow your audio.

Longest clip
~5MIN
Length follows your audio - roughly 5 minutes at the outside.
Resolutions
2
720p or 1080p; 1080p costs more per second.
Speech to lip sync
1:1
Delivered lines matched our source audio word for word in tests.
Speech or song
SING
Works from spoken lines or a sung performance.
Portrait needed
1
One clear, front-facing photo is enough to start.
Direction prompt
OPT
For Kling Avatar 2.0, an optional prompt shapes the presenter’s delivery and gestures while your recording supplies the actual words and timing.

Results

What you can make with Kling Avatar 2.0

  • From a portrait and a voice

    A nervous confession, timed to the audio

    A single portrait plus an audio clip of someone rehearsing what to say is enough: the avatar's glance down, the shy laugh at the end, and the length of the whole clip all follow the audio.

    Talking avatar, 1080p: dialogue matched the source audio word for word, expression followed the direction prompt.

  • Singing works too

    A performance, not just a reading

    Sent with a sung line rather than spoken dialogue, the avatar's hand gestures and gaze toward an imagined crowd followed the performance direction written into the prompt.

    Talking avatar, 720p: a between-songs address, gesture and gaze following the performance prompt.

  • Direction, in words

    One line of prompt sets the whole performance

    "Confident and upbeat, like opening a show, lively eyebrows and a bright smile" was the entire direction for this clip - the prompt field is where you set mood and gesture, separate from what the audio says.

    Talking avatar, 720p: an upbeat cold-open line delivered with matching eyebrow and smile animation.

Content test

Is Kling Avatar 2.0 uncensored? What we tested

Content levels tested on Kling Avatar 2.0, by mode
LevelTalking avatar
Mild
Not tested yet
Moderate
Not tested yet
Strong
RefusedTested Sep 28, 2026
Very strong
Not tested yet
Maximum
Not tested yet
How we test

How each mode behaves

Talking avatar
Clip length is set by your audio, not a duration field you choose directly - trim the audio to the length you want the clip to run. A clear, front-facing, well-lit portrait gives the most reliable lip sync and expression; side angles or obscured faces are more likely to drift.
All modes
Kling Avatar 2.0 reviews the portrait, voice-driven request and generated performance; adult NSFW content can trigger rejection even with an uploaded recording, and rejected requests are not charged. SpicyAPI adds no separate filter to Kling Avatar 2.0, and cannot disable Kling’s checks. MiniMax H3 Spicy is a separate unrestricted choice from Kling Avatar 2.0, with its own input options.
Talking avatar
There's no video-only mode: every request needs both a portrait and an audio file. There is no way to generate an avatar clip from text alone.

Use cases

What to build with the Kling Avatar 2.0 API

  • Turn a script and a voice recording into a face

    Record or generate the line you need spoken, pair it with a portrait, and get a clip of that person delivering it - useful for narrated intros, testimonials or character dialogue.

  • A singing performance from a still photo

    Pair a portrait with a sung audio track and direct the performance (gesture, gaze, energy) with a short prompt, as in the between-songs example on this page.

  • Consistent hosting across an episodic series

    Reuse the same portrait across episodes with new audio each time to keep a recurring host or character's face consistent from one clip to the next.

  • Multilingual delivery from the same face

    Because the performance follows whatever audio you send, the same portrait can deliver lines in different languages by swapping only the audio file.

Prompts

Kling Avatar 2.0 prompt examples you can copy

  • Talking avatar

    Direct the emotional arc, not just the mood

    Result from this prompt

    Nervous and sincere, glancing down once, then back into the lens with a shy laugh at the end.

    Why it works. Naming a sequence (glance down, then back up, then a laugh) rather than a single static mood gave the avatar a small arc to perform across the clip instead of one fixed expression held throughout.

    Try this prompt
  • Talking avatar

    Set the audience and the physical detail together

    Result from this prompt

    A warm, intimate delivery to a small late-night audience, a slight smile between lines, one hand resting on the microphone stand.

    Why it works. Naming who the performance is for (a small late-night audience) alongside a specific hand position gave the avatar a consistent gesture to hold rather than idle default motion.

    Try this prompt
  • Talking avatar

    Match energy to the line's content

    Result from this prompt

    Confident and upbeat, like opening a show, lively eyebrows and a bright smile.

    Why it works. A short, high-energy direction matched to an actual show-opening line kept the eyebrows and smile animated through the whole clip instead of flattening out partway through.

    Try this prompt
  • Talking avatar

    Write the relationship into the delivery

    Result from this prompt

    Relaxed and wry, as if talking to an old friend across the bar, small nods and an easy smile.

    Why it works. Framing the delivery around a specific relationship (an old friend across the bar) gave the avatar a consistent, low-key register rather than a generic neutral read.

    Try this prompt

Compare

Kling Avatar 2.0 vs similar models: which one to pick

ModelModesFromContentPick it when
Kling Avatar 2.0This pageTalking avatar$0.056 / output secondFiltered — This model refuses requests beyond its own limits outright. Measured, not assumed — models we have not measured carry no label at all.You have a portrait and an audio file and want a talking or singing clip whose length follows that audio.
Kling 2.6Character animation · Image to video · Text to video$0.069 / output secondFiltered — This model refuses requests beyond its own limits outright. Measured, not assumed — models we have not measured carry no label at all.You want a full scene with camera movement and setting, not just a portrait speaking, still with generated sound.
Kling 3.0Character animation · Image to video · Text to video$0.067 / output secondFiltered — This model refuses requests beyond its own limits outright. Measured, not assumed — models we have not measured carry no label at all.You want a full cinematic shot rather than a fixed-portrait performance, with 4K and Motion Control available.
Talking AvatarTalking avatar$0.018 / output second—You want to compare a different talking-avatar model for the same portrait-plus-audio workflow.

Pricing

Kling Avatar 2.0 API pricing

A 5-second clip starts at $0.28 at 720p and $0.56 at 1080p.

FAQ

Kling Avatar 2.0 API FAQ

About Kling Avatar 2.0

Kling Avatar 2.0 API animates a portrait to an uploaded voice or song: the audio supplies lip-sync timing and sets the clip length, with 720p or 1080p output. Kling Avatar 2.0 runs as an asynchronous task billed in US dollars per second of input audio; check the selected settings before submission, and retrieve the result by task ID or webhook. Kling Avatar 2.0 reviews the portrait, voice-driven request and generated performance; adult NSFW content can trigger rejection even with an uploaded recording, and rejected requests are not charged. SpicyAPI adds no separate filter to Kling Avatar 2.0, and cannot disable Kling’s checks. MiniMax H3 Spicy is a separate unrestricted choice from Kling Avatar 2.0, with its own input options. Before submitting to Kling Avatar 2.0, obtain the rights and consent for the source material and take responsibility for the resulting video.

Related models and next tasks