Verging AI API Documentation

Authentication

All API requests require an API key. Include it in the Authorization header:

Authorization: ApiKey vrg_sk_your_api_key_here

Generate API keys from your account dashboard at verging.ai (click your avatar → API Keys). The full key is shown only once at creation — store it securely.

Credits & Pricing

All services are billed using credits. Purchase credits at $10 = 530 credits.

Error Handling

All errors follow a consistent JSON format:

{
  "error_code": 40001,
  "message": "Insufficient credits. Required: 5, available: 2",
  "upstream_error": null
}

Rate Limits

AI Proxy services are rate-limited per user. When rate-limited, the response includes a 429 status code. Check the X-RateLimit-Reset header for when you can retry.

Swap faces in videos with AI precision. Supports HD processing and async job tracking. Try it online →

Create Job

Content-Type: multipart/form-data

curl -X POST https://verging.ai/api/v1/faceswap/create-job \
  -H "Authorization: ApiKey vrg_sk_xxx" \
  -F "swap_image=@face.jpg" \
  -F "file_name=video.mp4" \
  -F "target_video_url=https://cdn.verging.ai/uploads/video.mp4" \
  -F "user_video_duration=30" \
  -F "is_hd=false"

Response includes job_id. Poll the jobs endpoint for status updates.

List Jobs

Retry Failed Job

Only failed jobs can be retried. Credits are re-frozen for the retry.

Upscale video quality to high resolution using AI. Async job-based processing. Try it online →

Create Job

Content-Type: multipart/form-data

curl -X POST https://verging.ai/api/v1/video-enhance/create-job \
  -H "Authorization: ApiKey vrg_sk_xxx" \
  -F "file_name=video.mp4" \
  -F "target_video_url=https://cdn.verging.ai/uploads/video.mp4" \
  -F "user_video_duration=15"

Remove image backgrounds with AI. Fixed cost of 1 credit per image. Try it online →

Create Job

Content-Type: multipart/form-data

curl -X POST https://verging.ai/api/v1/background-removal/create-job \
  -H "Authorization: ApiKey vrg_sk_xxx" \
  -F "image=@photo.jpg" \
  -F "file_name=photo.jpg"

Generate AI images using DALL-E / gpt-image-1.

curl -X POST https://verging.ai/api/v1/ai/imagegen \
  -H "Authorization: ApiKey vrg_sk_xxx" \
  -H "Content-Type: application/json" \
  -d '{
    "prompt": "A futuristic cityscape at sunset, digital art style",
    "quality": "hd",
    "n": 2
  }'

Response

{
  "images": [
    "https://cdn.verging.ai/proxy/imagegen/abc123.png"
  ],
  "credits_consumed": 3
}

Generate chat completions using GPT-4o. Supports streaming (SSE) and non-streaming responses. Requires minimum 5 credits balance.

Non-streaming

curl -X POST https://verging.ai/api/v1/ai/chat \
  -H "Authorization: ApiKey vrg_sk_xxx" \
  -H "Content-Type: application/json" \
  -d '{
    "messages": [
      {"role": "system", "content": "You are a helpful assistant."},
      {"role": "user", "content": "Explain quantum computing in one paragraph."}
    ],
    "temperature": 0.7,
    "max_tokens": 256
  }'

Streaming (SSE)

curl -N -X POST https://verging.ai/api/v1/ai/chat \
  -H "Authorization: ApiKey vrg_sk_xxx" \
  -H "Content-Type: application/json" \
  -d '{
    "messages": [{"role": "user", "content": "Write a haiku about coding."}],
    "stream": true
  }'

Response (non-streaming)

{
  "content": "Quantum computing leverages quantum mechanical phenomena...",
  "usage": {
    "prompt_tokens": 24,
    "completion_tokens": 85,
    "total_tokens": 109
  },
  "credits_consumed": 1
}

Streaming responses use Server-Sent Events. Each chunk: data: {"content": "..."}. Final chunk includes usage and credits_consumed, followed by data: [DONE].

Convert text to speech audio using OpenAI TTS-1-HD.

curl -X POST https://verging.ai/api/v1/ai/tts \
  -H "Authorization: ApiKey vrg_sk_xxx" \
  -H "Content-Type: application/json" \
  -d '{
    "input": "Hello! Welcome to Verging AI.",
    "voice": "nova",
    "speed": 1.1
  }'

Response

{
  "audio_url": "https://cdn.verging.ai/proxy/tts/abc123.mp3",
  "credits_consumed": 1
}

Transcribe audio to text using OpenAI Whisper. Supports mp3, mp4, mpeg, mpga, m4a, wav, webm (max 25MB).

Content-Type: multipart/form-data

curl -X POST https://verging.ai/api/v1/ai/stt \
  -H "Authorization: ApiKey vrg_sk_xxx" \
  -F "file=@recording.mp3" \
  -F "language=en"

Response

{
  "text": "Hello, this is a sample transcription.",
  "language": "en",
  "duration": 12.5,
  "credits_consumed": 1
}

Analyze images using GPT-4o Vision. Supports image URL and file upload. Requires minimum 5 credits balance.

Option A — Image URL

curl -X POST https://verging.ai/api/v1/ai/vision \
  -H "Authorization: ApiKey vrg_sk_xxx" \
  -H "Content-Type: application/json" \
  -d '{
    "prompt": "Describe what you see in this image.",
    "image_url": "https://example.com/photo.jpg"
  }'

Option B — File Upload

curl -X POST https://verging.ai/api/v1/ai/vision/upload \
  -H "Authorization: ApiKey vrg_sk_xxx" \
  -F "prompt=What objects are in this image?" \
  -F "file=@photo.jpg"

Response

{
  "content": "The image shows a modern office space with several desks...",
  "usage": {
    "prompt_tokens": 1100,
    "completion_tokens": 95,
    "total_tokens": 1195
  },
  "credits_consumed": 2
}

Query your AI proxy service usage history.

curl https://verging.ai/api/v1/ai/usage \
  -H "Authorization: ApiKey vrg_sk_xxx"

Response

{
  "records": [
    {
      "id": 1,
      "service_type": "chat",
      "credits_consumed": 3,
      "input_tokens": 24,
      "output_tokens": 85,
      "latency_ms": 1200,
      "status": "success",
      "created_at": "2026-03-20T10:30:00"
    }
  ],
  "total": 17,
  "page": 1,
  "page_size": 20
}

Manage your API keys. Create new keys from your account dashboard at verging.ai (avatar → API Keys). Keys use the prefix vrg_sk_.

List API Keys

Returns your API keys (prefix only, never the full key).

Delete API Key

Permanently revokes the specified API key.

Validate API Key

Test if your API key is valid and active.

curl -X POST https://verging.ai/api/v1/api-keys/validate \
  -H "Authorization: ApiKey vrg_sk_xxx"

Use Verging AI directly from your terminal with agent skills. Each skill wraps our API into a single command — no code needed. Skills work with Claude Code and compatible AI coding agents.

Prerequisites

export VERGING_API_KEY="vrg_sk_your_key_here"

Install a Skill

Skills are installed via the OpenClaw registry. Each skill is a single command:

# Install face swap skill
npx openclaw add verging-ai/faceswap

# Install video enhancement skill
npx openclaw add verging-ai/video-enhancement

# Install background removal skill
npx openclaw add verging-ai/background-removal

All 8 skills are available at github.com/verging-ai/agent-skills.

Swap faces in videos from your terminal. Supports local files, YouTube/Bilibili URLs, time trimming, and HD mode.

# Basic usage — local video + local face image
/faceswap -v ./input.mp4 -f ./my-face.jpg

# Remote video from YouTube
/faceswap -v "https://youtube.com/watch?v=xxx" -f ./face.jpg --hd

# Trim to specific time range (saves credits)
/faceswap -v ./video.mp4 -f ./face.jpg --start 5 --end 15

# Auto-download result to current directory
/faceswap -v ./video.mp4 -f ./face.jpg --download

Requires yt-dlp for remote video URLs. ffmpeg for time trimming.

Upscale and enhance video quality from your terminal. Restore old footage, sharpen low-res videos, upscale to HD/4K.

# Enhance a local video
/video-enhancement -v ./old-video.mp4

# HD mode for maximum quality
/video-enhancement -v ./clip.mp4 --hd

# Enhance a specific segment from YouTube
/video-enhancement -v "https://youtube.com/watch?v=xxx" --start 5 --end 15 --hd

# Download result automatically
/video-enhancement -v ./video.mp4 --download

Normal mode: 1 credit/sec. HD mode: 3 credits/sec. Max 30 seconds per video.

Remove image backgrounds with one command. Outputs transparent PNG. Great for product photos, headshots, and design assets.

# Remove background from local image
/background-removal -i ./photo.jpg

# From a remote URL
/background-removal -i "https://example.com/product.jpg"

# Save to specific path
/background-removal -i ./photo.jpg -o ./transparent.png

Fixed cost: 1 credit per image. Supports JPG, PNG, WebP (max 10MB).

In addition to the core video and image skills above, Verging AI offers CLI skills for all AI proxy services. These use the same API key and credit system.

All skills follow the same pattern: install via npx openclaw add verging-ai/<skill-name>, set VERGING_API_KEY, and run the command. Full documentation for each skill is available in the GitHub repository.