Claude visitors: new accounts receive $1.00 in bonus API credit within 30 days. One per person; usage credits only.
Use InfiniteTalk in WorkBuddy.
InfiniteTalk generates lip-synced talking avatar videos from an audio file and a face image. Add RunAPI as an MCP server in WorkBuddy, prompt the workbench agent with the audio and image URLs, and the InfiniteTalk tool returns a finished video URL — the async polling happens behind the scenes.
Generate a talking avatar video with InfiniteTalk using the runapi MCP server.
Requirements:
- Use the RunAPI MCP tool for audio-to-video generation.
- Set model to "infinitetalk-from-audio".
- Set source_audio_url to https://example.com/speech.mp3.
- Set source_image_url to https://example.com/avatar.jpg.
- Prompt: "A professional presenter delivering a product overview".
- The task is async — wait until it completes, then give me the video URL.
{
"mcp_servers": {
"runapi": {
"url": "https://runapi.ai/mcp",
"transport": "streamable_http",
"headers": {
"Authorization": "Bearer $RUNAPI_API_KEY"
}
}
}
}
{
"task_id": "tsk_abc123",
"status": "completed",
"model": "infinitetalk-from-audio",
"output": {
"video_url": "https://media.runapi.ai/outputs/tsk_abc123.mp4"
}
}
Use InfiniteTalk in WorkBuddy in three steps
Add the RunAPI MCP server
Register an MCP server named runapi in WorkBuddy pointing at https://runapi.ai/mcp over streamable HTTP, with your key read from the RUNAPI_API_KEY environment variable. If RunAPI is already connected for chat, the same server exposes InfiniteTalk — no extra setup.
export RUNAPI_API_KEY=runapi_xxx
# MCP server: runapi → https://runapi.ai/mcp
Prompt the workbench agent
Give WorkBuddy the audio URL, the face image URL, and a short description of the desired output. The agent calls the InfiniteTalk audio-to-video tool with model set to infinitetalk-from-audio and fills in source_audio_url, source_image_url, and prompt.
"Turn speech.mp3 + avatar.jpg into a talking avatar video"
Let the tool return the result
The RunAPI MCP tool creates the task, waits through the async generation until the status is completed, and hands the lip-synced video URL back into the workbench — no manual polling needed.
status: completed → video_url
InfiniteTalk parameters
| Parameter | Type | Description |
|---|---|---|
model |
string |
Required. Use infinitetalk-from-audio. |
source_audio_url |
string |
Required. URL of the source audio file to lip-sync. |
source_image_url |
string |
Required. URL of the face or avatar image to animate. |
prompt |
string |
Required. Text description of the desired video output. Max 5000 characters. |
output_resolution |
string |
Optional. Output video resolution. Accepted values: 480p, 720p. Defaults to 480p. |
seed |
integer |
Optional. Reproducibility seed. Integer between 10000 and 1000000. |
What is InfiniteTalk on WorkBuddy?
InfiniteTalk is an audio-driven talking avatar model that turns a single photo and an audio file into a lip-synced video with head movement, facial expressions, and body posture. Through the runapi MCP server, WorkBuddy agents call it as a standard tool to produce unlimited-length talking head videos without filming equipment or motion capture rigs.
InfiniteTalk use cases
Podcast-to-video conversion
Feed podcast audio and a host photo to generate a talking avatar video, turning audio-only episodes into visual content for YouTube or social media.
Educational and tutorial videos
Create lecture or tutorial videos from a single instructor photo and recorded narration, producing long-form talking head content without a camera setup.
Digital spokesperson for marketing
Build a branded digital presenter from one reference image, then generate spokesperson videos in multiple languages by swapping the audio track.
InfiniteTalk + WorkBuddy questions
Yes. WorkBuddy speaks MCP natively, so add RunAPI as an MCP server named runapi with your RUNAPI_API_KEY, then prompt the agent with an audio URL and a face image URL. The InfiniteTalk tool handles the generation and returns the video.
Yes. InfiniteTalk uses a sparse-frame approach that supports unlimited-length video generation from one reference image. The output length is determined by the duration of your source audio file.
No. The same RUNAPI_API_KEY behind the runapi MCP server covers InfiniteTalk along with every other RunAPI model. All usage draws from one pooled credit balance.
Asynchronous. The task returns a task_id immediately, and the RunAPI MCP tool keeps checking the status until the video is ready — WorkBuddy just receives the finished URL, so you never poll by hand.
InfiniteTalk uses per-task billing. Check the RunAPI pricing page for current rates. Credits are deducted from the same balance used for all RunAPI models.
WorkBuddy general setup
Not connected yet? Start with the RunAPI MCP setup guide for WorkBuddy.
WorkBuddy setup guide →InfiniteTalk model catalog
See InfiniteTalk variants, pricing, and API docs.
InfiniteTalk on RunAPI →Try InfiniteTalk in WorkBuddy today.
Get a free RunAPI key, add the MCP server to WorkBuddy, and generate lip-synced talking avatar videos from any audio file.