> ## Documentation Index > Fetch the complete documentation index at: https://docs.goenhance.ai/llms.txt > Use this file to discover all available pages before exploring further. # AI Talking Avatar (Video) > AI Talking Avatar from a video. **How it works** — one video of a person plus one audio track: the person in the video speaks the audio. The whole frame is regenerated with your video as the reference, so head movement and facial expressions follow the speech — not just the lips. To only redraw the mouth and keep every other pixel of your video, use `lipsync-video` instead; to start from a single photo, use `ai-talking-avatar`. **Output length** — set by `video_mode` when the video and the audio have different lengths: | video_mode | output length | |---|---| | `normal` (default) | the shorter of the video and the audio | | `loop` | the audio length; a shorter video is played forward and backward until it covers the audio | Before any tokens are deducted, GoEnhance transfers both files to its own storage and measures their real durations — that measurement is what you are billed for, and it is also what enforces the 3-60s limit. Because the files are measured and then sent onward from GoEnhance's storage, swapping the URLs afterwards has no effect. To make a shorter video, set the optional `duration` (whole seconds, 3-60): only the first `duration` seconds are used, and you are billed for `duration` seconds. It cannot exceed the output length above. With `duration` set, the video and the audio themselves may be longer than 60 seconds. **Video requirements** — up to 300MB. One person, face clearly visible and roughly front-facing for most of the clip. The output keeps the framing of your video. **Audio requirements** — up to 50MB. Clear speech with little background noise gives the best lip sync. **Pricing** — per second of output, in tokens (USD at $0.02/token): | resolution | tokens/s | USD/s | |---|---|---| | 540p | 1 | $0.02 | | 720p | 1.5 | $0.03 | A 30-second clip therefore costs 30 tokens at 540p (= $0.60). At 720p the same clip costs 45 tokens (= $0.90). Returns an `img_uuid`; poll GET /api/v1/jobs/detail (or use `custom_callback_url`) to get the generated video. ## OpenAPI ````yaml api-reference/video-generations/openapi-ai-talking-avatar-video.json POST /api/v1/videos/generations openapi: 3.1.0 info: title: GoEnhance API - AI Talking Avatar (Video) description: >- Make the person in a video speak an audio track, with head movement and expressions that follow the speech. version: 1.0.0 servers: - url: https://api.goenhance.ai security: [] paths: /api/v1/videos/generations: post: tags: - VideoGenerations summary: AI Talking Avatar (Video) description: >- AI Talking Avatar from a video. **How it works** — one video of a person plus one audio track: the person in the video speaks the audio. The whole frame is regenerated with your video as the reference, so head movement and facial expressions follow the speech — not just the lips. To only redraw the mouth and keep every other pixel of your video, use `lipsync-video` instead; to start from a single photo, use `ai-talking-avatar`. **Output length** — set by `video_mode` when the video and the audio have different lengths: | video_mode | output length | |---|---| | `normal` (default) | the shorter of the video and the audio | | `loop` | the audio length; a shorter video is played forward and backward until it covers the audio | Before any tokens are deducted, GoEnhance transfers both files to its own storage and measures their real durations — that measurement is what you are billed for, and it is also what enforces the 3-60s limit. Because the files are measured and then sent onward from GoEnhance's storage, swapping the URLs afterwards has no effect. To make a shorter video, set the optional `duration` (whole seconds, 3-60): only the first `duration` seconds are used, and you are billed for `duration` seconds. It cannot exceed the output length above. With `duration` set, the video and the audio themselves may be longer than 60 seconds. **Video requirements** — up to 300MB. One person, face clearly visible and roughly front-facing for most of the clip. The output keeps the framing of your video. **Audio requirements** — up to 50MB. Clear speech with little background noise gives the best lip sync. **Pricing** — per second of output, in tokens (USD at $0.02/token): | resolution | tokens/s | USD/s | |---|---|---| | 540p | 1 | $0.02 | | 720p | 1.5 | $0.03 | A 30-second clip therefore costs 30 tokens at 540p (= $0.60). At 720p the same clip costs 45 tokens (= $0.90). Returns an `img_uuid`; poll GET /api/v1/jobs/detail (or use `custom_callback_url`) to get the generated video. parameters: - name: Authorization in: header description: '' required: false example: '{{Authorization}}' schema: type: string requestBody: content: application/json: schema: type: object properties: model: type: string enum: - ai-talking-avatar-video description: Model name. Must be `ai-talking-avatar-video`. video_url: type: string format: uri description: >- Video of the person who should speak (up to 300MB). Required. Keep one person in frame, with the face clearly visible for most of the clip. audio_url: type: string format: uri description: Speech audio to lip-sync to (up to 50MB). Required. prompt: type: string default: person talking description: >- Optional description of the desired performance, e.g. `a man speaking calmly to the camera`. resolution: type: string enum: - 540p - 720p default: 540p description: >- Output resolution. 720p costs 1.5 times as much per second as 540p. video_mode: type: string enum: - normal - loop default: normal description: >- What to do when the video and the audio have different lengths. `normal`: the output is as long as the shorter of the two. `loop`: the output follows the audio, and a shorter video is played forward and backward until it covers the audio. duration: type: integer minimum: 3 maximum: 60 description: >- Optional. Output length in whole seconds (3-60). Only the first `duration` seconds are used, and you are billed for `duration` seconds. Must not exceed the output length set by `video_mode`. Omit it to use the full length. example: 15 custom_callback_url: type: string format: uri description: >- Optional. A publicly accessible HTTPS URL. When the task status changes (processing / success / failed), GoEnhance sends a POST request to this URL. The request body is identical to the response of GET /api/v1/jobs/detail. If your server does not respond with HTTP 200, the notification is retried up to 3 times, with a 3-second timeout per attempt. example: https://your-server.com/goenhance/callback required: - model - video_url - audio_url example: model: ai-talking-avatar-video video_url: https://example.com/speaker.mp4 audio_url: https://example.com/speech.mp3 prompt: a man speaking calmly to the camera resolution: 540p video_mode: normal responses: '200': description: '' content: application/json: schema: type: object properties: code: type: integer msg: type: string data: type: object properties: img_uuid: type: string cost: type: number description: >- Tokens deducted for this request. This is the amount actually charged, so it already reflects any discount active on your account and can be lower than the listed price. Tokens are deducted when the task is accepted, and refunded automatically if the generation ends in failure. example: 5.82 required: - img_uuid - cost required: - code - msg - data examples: '1': summary: Success value: code: 0 msg: Success data: img_uuid: c12b656c-747a-44fd-9c80-add79b0c52d5 cost: 5.82 '2': summary: Insufficient tokens value: code: 100 msg: tokens is not enough quota: summary: API key quota exceeded value: code: 101 msg: 'API key quota exceeded: 4990 of 5000 tokens used (monthly)' headers: {} '401': description: '' content: application/json: schema: type: object properties: code: type: integer msg: type: string required: - code - msg headers: {} deprecated: false ```` This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.