Integration

API4AI integration

API4AI is a cloud computer vision platform. Send an image (as a file or a public URL) to detect NSFW content, read text (OCR) and VINs, label images, detect objects, brand logos, faces and fashion items, recognize wine and alcohol labels and household items, remove or replace backgrounds (general, people, cars), anonymize faces and license plates, upscale images, or dress a person in an apparel photo (virtual try-on). Results come back as JSON; result images are base64 inside it. Service version, warm-up and the list-only helper routes (algorithms, modes, kinds) are not exposed; their values are documented on the parameters. Authentication is an API4AI Developer Portal key (portal.api4.ai, prepaid pay-as-you-go) sent in the X-API-KEY header.

Connect API4AI How integrations work

Ask it in chat

Say what should happen.
The workflow builds itself.

When this happens, do that in API4AI.Built from the 17 API4AI tools below, on the connection you add once.

What you can do with API4AI

Each of these is one operation on one endpoint, available both as a workflow step and as a tool an AI agent can call on its own. Nothing bundles several operations into one action, so what a run did stays readable afterwards.

Most endpoints run on the free plan. A few, mainly the ones that publish on your behalf, need a paid plan.

analyze_facesAnalyze facesPOSTDetects faces in an image and returns, per face, a bounding box, a confidence, five facial landmarks (eyes, nose tip, mouth corners) and, with embeddings=true, a 512-value embedding vector for face matching.
anonymize_imageAnonymize imagePOSTHides faces and car license plates in an image (or only the kinds given in mode) and returns the anonymized image as base64 plus the boxes of the hidden objects.
detect_brandsDetect brandsPOSTFinds brand logos in an image. By default returns the list of brand names found; with detailed=true each logo instance with its size category, relative bounding box and the object it is printed on.
detect_nsfwDetect nsfwPOSTChecks an image (or each PDF page) for sexual content and returns the confidence of the nsfw and sfw classes (0.0 to 1.0). Lower strictness to tolerate borderline images.
detect_objectsDetect objectsPOSTDetects general objects in an image and returns each with a relative bounding box and recognized classes with confidence. Choose the algorithm with algo.
label_imageLabel imagePOSTClassifies an image into general categories and returns each predicted label with its probability (0.0 to 1.0). Choose the algorithm with algo.
recognize_alcohol_labelRecognize alcohol labelPOSTReads a bottle or can label and returns the drink kind (beer, wine, vodka, whiskey, bourbon, brandy, cognac, rum, tekila or liqueur) with details that vary by drink, such as country, variety, vintage and abv for wine or brewery and type for beer.
recognize_fashion_itemsRecognize fashion itemsPOSTDetects clothing and fashion accessories in an image and returns each item with its bounding box and recognized classes with confidence scores.
recognize_household_itemsRecognize household itemsPOSTRecognizes furniture and household items in a room photo and returns a count per item name. Choose the item set for moving or for cleaning companies with kind.
recognize_textRecognize textPOSTReads the text in an image or each page of a PDF (OCR) and returns text blocks with their bounding boxes: the whole text as one block (simple-text, default) or each word separately (simple-words).
recognize_vinRecognize vinPOSTReads a vehicle identification number (VIN) from an image, applying VIN rules (character set, length, checksum), and returns one or more suggested VINs.
recognize_wineRecognize winePOSTRecognizes a wine from a photo of its bottle label and returns the top n matching wines (name and vintage) with their confidence.
remove_backgroundRemove backgroundPOSTRemoves the background of a general image and returns a transparent PNG of the foreground (or its mask with mode fg-mask) as base64. An optional background image can be blended in.
remove_background_carsRemove background carsPOSTRemoves the background of a car photo and returns the car as a transparent PNG (base64), optionally with a realistic shadow and the license plate hidden, or the mask. An optional background image can be blended in.
remove_background_peopleRemove background peoplePOSTRemoves the background behind people in a photo and returns a transparent PNG of the person (or the mask with mode fg-mask) as base64. An optional background image can be blended in.
upscale_imageUpscale imagePOSTUpscales an image (input up to 1024x1024) and returns the larger image as a base64-encoded PNG or JPEG in the result entities.
virtual_try_onVirtual try onPOSTDresses the person of one image in the apparel of a second image and returns the generated photo as a base64 PNG. An optional prompt can adjust the scene, such as the background or time of day.

Use API4AI from Claude Code, Cursor or Codex

Connect API4AI in LiveContext, create an API key in Settings, then add the LiveContext MCP server to your client. Its API4AI operations become tools the client can call, next to every other integration you connected.

Claude Code

claude mcp add --transport http livecontext https://livecontext.ai/mcp --header "X-API-Key: YOUR_API_KEY"

Codex (~/.codex/config.toml)

[mcp_servers.livecontext]
url = "https://livecontext.ai/mcp"

[mcp_servers.livecontext.http_headers]
"X-API-Key" = "YOUR_API_KEY"

Cursor, Claude Desktop and other JSON configs

{
  "mcpServers": {
    "livecontext": {
      "type": "http",
      "url": "https://livecontext.ai/mcp",
      "headers": {
        "X-API-Key": "YOUR_API_KEY"
      }
    }
  }
}

Set up the MCP server step by step

How to connect API4AI

Create your workspace
Sign up free. Nothing to install.
Connect API4AI once
Add your API4AI connection in Settings. Every tool from this integration reuses it, and secrets are encrypted at rest and visible only to you.
Describe the job in chat
The workflow builds itself in front of you and picks the endpoints it needs. Drag the steps yourself on the canvas whenever you would rather be explicit.