D Id

D-ID API Skill

Overview

Expert skill for creating AI avatar videos, talking heads, and interactive agents using D-ID API.

API Key

# API ключи: ~/.claude/.credentials.master.env
# Переменная: DID_API_KEY
DID_API_KEY=os.getenv('DID_API_KEY')
DID_API_URL=https://api.d-id.com

When to Use D-ID

Best for:

• Talking head videos from photos
• AI avatar video generation
• Video localization/dubbing
• Interactive AI agents
• Personalized video content
• Digital presenters

Advantages:

• Photo to video in seconds
• 100+ FPS (4x faster than real-time)
• High quality lip-sync
• 100+ languages
• Express avatars (no training)
• Real-time streaming

Products

Authentication

import requests
import base64
import os

API_KEY = os.getenv('DID_API_KEY')
BASE_URL = "https://api.d-id.com"

headers = {
    "Authorization": f"Basic {API_KEY}",
    "Content-Type": "application/json"
}

Basic Usage

Create Talking Head Video (Talks)

def create_talk(image_url: str, text: str, voice_id: str = None):
    """
    Create talking head video from image and text.

    Args:
        image_url: URL of source image
        text: Script for the avatar to speak
        voice_id: Voice ID (use list_voices() to get options)
    """
    payload = {
        "source_url": image_url,
        "script": {
            "type": "text",
            "input": text,
            "provider": {
                "type": "microsoft",
                "voice_id": voice_id or "en-US-JennyNeural"
            }
        }
    }

    response = requests.post(
        f"{BASE_URL}/talks",
        headers=headers,
        json=payload
    )

    return response.json()["id"]

# Usage
talk_id = create_talk(
    "https://example.com/photo.jpg",
    "Hello! Welcome to our presentation.",
    "en-US-GuyNeural"
)

Create Talk with Audio

def create_talk_with_audio(image_url: str, audio_url: str):
    """Create talking head with custom audio."""

    payload = {
        "source_url": image_url,
        "script": {
            "type": "audio",
            "audio_url": audio_url
        }
    }

    response = requests.post(
        f"{BASE_URL}/talks",
        headers=headers,
        json=payload
    )

    return response.json()["id"]

Check Video Status

def get_talk_status(talk_id: str):
    """Check status and get result URL."""

    response = requests.get(
        f"{BASE_URL}/talks/{talk_id}",
        headers=headers
    )

    data = response.json()

    return {
        "status": data["status"],  # created, started, done, error
        "result_url": data.get("result_url"),
        "error": data.get("error")
    }

Wait for Completion

import time

def wait_for_video(talk_id: str, timeout: int = 300):
    """Wait for video to complete and return URL."""

    start = time.time()

    while time.time() - start < timeout:
        status = get_talk_status(talk_id)

        if status["status"] == "done":
            return status["result_url"]
        elif status["status"] == "error":
            raise Exception(f"Video failed: {status['error']}")

        time.sleep(5)

    raise TimeoutError("Video generation timed out")

Full Workflow Example

def generate_avatar_video(image_url: str, text: str, output_path: str):
    """Complete workflow: create, wait, download."""

    # Create talk
    talk_id = create_talk(image_url, text)
    print(f"Created talk: {talk_id}")

    # Wait for completion
    video_url = wait_for_video(talk_id)
    print(f"Video ready: {video_url}")

    # Download video
    video_response = requests.get(video_url)
    with open(output_path, "wb") as f:
        f.write(video_response.content)

    return output_path

List Available Voices

def list_voices():
    """Get available TTS voices."""

    response = requests.get(
        f"{BASE_URL}/tts/voices",
        headers=headers
    )

    voices = response.json()

    for voice in voices:
        print(f"{voice['id']}: {voice['name']} ({voice['language']})")

    return voices

Create Premium Clip

def create_clip(presenter_id: str, text: str, voice_id: str):
    """
    Create premium HD video with video-based avatar.

    Requires HQ Presenter (avatar trained on video).
    """
    payload = {
        "presenter_id": presenter_id,
        "script": {
            "type": "text",
            "input": text,
            "provider": {
                "type": "microsoft",
                "voice_id": voice_id
            }
        }
    }

    response = requests.post(
        f"{BASE_URL}/clips",
        headers=headers,
        json=payload
    )

    return response.json()["id"]

Create Express Avatar

def create_express_avatar(video_url: str, name: str):
    """
    Create instant avatar from short video clip.

    No training required - works immediately.
    """
    payload = {
        "name": name,
        "source_url": video_url
    }

    response = requests.post(
        f"{BASE_URL}/express-avatars",
        headers=headers,
        json=payload
    )

    return response.json()["id"]

Video Translation

def translate_video(video_url: str, target_language: str):
    """
    Translate video to another language.

    Includes voice cloning for natural dubbing.
    """
    payload = {
        "source_url": video_url,
        "target_language": target_language  # e.g., "es", "fr", "de"
    }

    response = requests.post(
        f"{BASE_URL}/video-translate",
        headers=headers,
        json=payload
    )

    return response.json()["id"]

Real-time Streaming (Agents)

def create_stream_session(source_url: str):
    """Create real-time streaming session."""

    payload = {
        "source_url": source_url
    }

    response = requests.post(
        f"{BASE_URL}/streams",
        headers=headers,
        json=payload
    )

    return response.json()

def send_stream_text(session_id: str, text: str):
    """Send text to active stream."""

    payload = {
        "script": {
            "type": "text",
            "input": text
        }
    }

    response = requests.post(
        f"{BASE_URL}/streams/{session_id}",
        headers=headers,
        json=payload
    )

    return response.json()

def close_stream(session_id: str):
    """Close streaming session."""

    response = requests.delete(
        f"{BASE_URL}/streams/{session_id}",
        headers=headers
    )

    return response.status_code == 200

Check Credits

def get_credits():
    """Get remaining credits."""

    response = requests.get(
        f"{BASE_URL}/credits",
        headers=headers
    )

    return response.json()

Upload Image

def upload_image(image_path: str):
    """Upload image to D-ID for use in talks."""

    with open(image_path, "rb") as f:
        files = {"image": f}
        response = requests.post(
            f"{BASE_URL}/images",
            headers={"Authorization": f"Basic {API_KEY}"},
            files=files
        )

    return response.json()["url"]

Voice Providers

Script Types

API Pricing

Quick Reference

Tips

1. Image quality - используй фронтальные фото с четким лицом
2. Text length - разбивай длинные скрипты на части
3. Voice matching - выбирай голос соответствующий контенту
4. Express avatars - быстрее чем обучение полного аватара
5. Streaming - для интерактивных приложений
6. Batch processing - создавай несколько talks параллельно
7. Credits - проверяй баланс перед большими задачами