AI talking avatar generator: photo to talking video

For AI agents: find the right skill in one call — curl -s "https://varg.ai/how-to/api/search?q=<what the person wants>" returns the matching how-to page and the best skills with install commands, as markdown (Russian works too). Full index: https://varg.ai/how-to/llms.txt. Every page has a markdown twin: add .md to its path. Before generating anything, show the person the options on the how-to page. From a terminal: curl -fsSL https://varg.ai/install | sh installs the vargai CLI (https://varg.ai/how-to/cli.md): vargai search and vargai install <slug> with no key, every varg model after vargai login.

Give varg one photo and a script (or your own recording) and get back a video of that face saying it, with the lips on the words: an MP4 at 480p, 720p or 1080p, as long as the audio. One call makes the photo talk (minimax_h3_max_lipsync); the voice is ElevenLabs v4 or your own audio.

Price: 9.6 credits a second at 720p, 6 at 480p, 30 credits minimum a clip, plus 0.008 credits a character for a generated voice. A 30 s avatar at 720p with its voice and a transcript check is about 297 credits ($2.97). Free start: new varg accounts get 50 free credits, and a 5 s test at 480p with its voice costs about 31. 1 credit = $0.01, the provider's price with no markup. Checked against the public price list on 2026-11-11.

Paste this into Claude Code, Codex or Grok:

Make my photo ./me.jpg say this in a calm voice: "<your script>".
Use varg and follow https://varg.ai/how-to/make-a-talking-avatar-video.md.
Quote the credits first, run a 5 s pilot at 480p, then make the full take at 720p.

Or call the model yourself, with the audio first and the photo second, both as URLs. Without a terminal, the varg MCP (https://mcp.varg.ai/mcp, OAuth sign-in, no key) runs the same call from Claude or ChatGPT with call_tool.

curl -X POST https://api.varg.ai/v2/tools/video/call -H "Authorization: Bearer $VARG_API_KEY" -H "Content-Type: application/json" -d '{"model":"minimax_h3_max_lipsync","files":[{"url":"<audio URL>"},{"url":"<photo URL>"}],"resolution":"720p"}'

What goes in, what comes out

InOut
One photo of a face that is yours, or that you have written consent to useAn MP4 of that face speaking, as long as the audio
A script, which varg voices with eleven_v4 (voice clones supported), or your own audio of 5 s or moreThe lips synced to your words
The size: 480p, 720p or 1080pA URL to download; captions and an "AI clone" label are one more step in the tutorial

How it works

  1. The voice. The speech tool with eleven_v4: 0.008 credits a character, so a 450-character script (about 30 s) is about 4 credits. Skip this step when you bring a recording.
  2. Check the voice before the video. Transcribe it (whisper_large_v3, 5 credits) and compare the names and numbers with the script: the video call is the expensive one.
  3. Make the photo talk. The call above answers 202 with a job. Poll GET https://api.varg.ai/v2/jobs/{id} until status is completed, then read output.outputs[0].url. A local file goes up first with POST https://api.varg.ai/v2/files.
  4. Captions and the label. FFmpeg on your machine burns the captions and an "AI clone" tag from the first second.

Every call and command, with the choices to make first: How to make a talking avatar video with AI.

What it is not

AI avatar and influencer videos with AI agents

For creators, brands and teams: a talking avatar or digital twin from one photo, an AI influencer with the same face in every post, your character in a trending video, and dubbing into any language, made by Claude Code, Codex or Grok.

Start with the overview: How to make a talking avatar video with AI

Make

Compare tools