Kling API avatar: one image plus audio, 2 to 300 seconds

Yes, Kling's API has an avatar endpoint: POST /v1/videos/avatar/image2video animates one reference image into a talking video driven by an audio track of 2 to 300 seconds, either a sound file or audio from Kling's own TTS. It runs in std or pro mode and is billed per second.

Kling's facts come from its Avatar, Video pricing and Authentication API pages, read on 2026-09-28. Sume does not route Kling's Avatar model; the Sume options further down come from the Sume API reference and the Generate avatar video docs.

What does the Kling Avatar API take?

The request is JSON with a Bearer API key from the Kling AI console. The result is a task you query or get a callback_url notification for.

How long do Kling avatar videos stay available?

Kling's response notes that generated videos "will be cleared after 30 days", so copy the file from task_result.videos[].url once the task reaches succeed. Kling's pricing table does not say which mode maps to which price column; the task result reports the units deducted in final_unit_deduction.

Can I make the same image-plus-audio video on Sume?

Yes, with a different model: POST /v1/veed/fabric-1.0 (VEED Fabric 1.0) turns a still image, or a ready avatar's still, plus audio into a talking clip. It costs $0.1875 per audio second (720p) on API pricing, and Lip sync API walks through the request. Where its inputs differ from Kling's Avatar call:

• Picture: exactly one visual source, a public HTTPS image_url or a ready avatar's avatar_id / avatar_handle.
• Audio: Kling takes a sound_file or a Kling TTS audio_id. Sume's audio_url must be a file on Sume's media host, typically a TTS segment, of at most 10 MB; other hosts are refused.
• Length and size: the audio's length goes in duration_seconds (1 to 300), and resolution is 480p or 720p (default 720p).

What if I only have a script, not audio?

Kling's route takes TTS audio from its own TTS endpoint. On Sume, the script-in route is Avatar 1.0, POST /v1/avatar-1.0/talking-video, which turns a ready avatar and a script into a talking video.

• Scripts are accepted when Sume estimates the video at 4 to 60 seconds; split longer scripts into several jobs.
• aspect_ratio is 1:1, 3:4, 9:16 (default), 4:3 or 16:9, and 720p is the documented resolution.
• In current code the avatar speaks English only. For other languages, generate speech with TTS 1.0 (POST /v1/tts-1.0/generate, with language set) and lip-sync it with Fabric.