---
title: "Use InfiniteTalk in Hermes Agent via RunAPI — Talking Avatar API Guide"
url: "https://runapi.ai/hermes-infinitetalk.md"
canonical: "https://runapi.ai/hermes-infinitetalk"
locale: "en"
model: "infinitetalk"
---

# Use InfiniteTalk in Hermes Agent.

InfiniteTalk generates lip-synced talking avatar videos from an audio file and a face image. Hermes Agent calls it through the RunAPI custom endpoint — configure custom:runapi once, send audio_to_video, and poll for the finished video.

## API example

```bash
curl -X POST https://runapi.ai/api/v1/infinitetalk/audio_to_video \
  -H "Authorization: Bearer $RUNAPI_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "infinitetalk-from-audio",
    "source_audio_url": "https://example.com/speech.mp3",
    "source_image_url": "https://example.com/avatar.jpg",
    "prompt": "A professional presenter delivering a product overview"
  }'

```

### Response

```json
{
  "task_id": "tsk_abc123",
  "status": "pending",
  "model": "infinitetalk-from-audio"
}

```

## How it works

1. **Configure RunAPI** — Set RUNAPI_API_KEY in the environment where Hermes Agent runs. If you already added RunAPI as a custom:runapi provider, the same key handles InfiniteTalk requests.
2. **Call InfiniteTalk** — Send a POST request to the audio_to_video endpoint with the model, audio URL, image URL, and a prompt. Hermes Agent can build and send this request through the custom:runapi provider.
3. **Poll for the result** — The endpoint returns a task_id immediately. Poll the task status endpoint until the status is completed, then read the output video URL from the response.

## 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. |
| `callback_url` | `string` | Optional. Webhook URL that receives a POST when the task completes. |

## FAQ

### Can I use InfiniteTalk in Hermes Agent?

Yes. Hermes Agent can call the RunAPI InfiniteTalk endpoint through the custom:runapi provider. Set RUNAPI_API_KEY in your environment and send a POST to /api/v1/infinitetalk/audio_to_video with the model, audio URL, and image URL.

### Can InfiniteTalk generate unlimited-length talking videos from a single photo?

Yes. InfiniteTalk uses a sparse-frame approach that supports unlimited-length video generation from one reference image. The output length matches the duration of the source audio file you provide.

### How do I route InfiniteTalk requests through the Hermes custom provider?

Add a custom:runapi provider in your Hermes Agent configuration with base_url set to https://runapi.ai/v1 and your RUNAPI_API_KEY. Hermes Agent then routes InfiniteTalk requests through this provider automatically.

### Can I chain InfiniteTalk with other RunAPI models in a Hermes Agent workflow?

Yes. Hermes Agent supports multi-step workflows. You can chain a TTS model like ElevenLabs to generate audio, then pass the audio URL to InfiniteTalk for avatar video generation, all within a single agent run.


## Links

- [Hermes Agent setup guide →](https://runapi.ai/hermes-agent)
- [InfiniteTalk on RunAPI →](https://runapi.ai/models/infinitetalk)
- [Model catalog](https://runapi.ai/models)
- [API docs](https://runapi.ai/docs)