# BLIP - Core.Today AI API

> Salesforce BLIP (173M+ runs) — image captioning, visual question answering, and image-text matching in one model. The classic choice for bulk captioning at 1 credit per image.

- **Provider**: Salesforce
- **Model ID**: salesforce/blip
- **Category**: Image Generation
- **Credits**: 1 per image
- **Speed**: Fast
- **Quality**: Standard

## Features
- Three tasks: captioning, VQA, image-text matching
- 173M+ runs — the classic captioner
- 1 credit per image — ideal for bulk pipelines
- Short, reliable English captions

## Use Cases
- Alt-text generation at scale
- Image library auto-tagging
- Dataset caption bootstrapping
- Simple visual QA
- Image-text relevance scoring

## API Endpoint

Base URL: https://api.core.today/v1
Create Prediction: POST /predictions
Get Status: GET /predictions/{job_id}

## Authentication
Header: X-API-Key: YOUR_API_KEY

## Input Parameters

### Required
- **image**: string - Input image

### Optional
- **task**: string (default: image_captioning) - Task to perform
  Options: image_captioning, visual_question_answering, image_text_matching
- **question**: string - Question about the image (VQA task only)
- **caption**: string - Caption to match against the image (matching task only)

## Examples

### Bulk Captioning
Generate a caption for an image
```json
{
  "model": "salesforce/blip",
  "input": {
    "image": "https://example.com/photo.jpg",
    "task": "image_captioning"
  }
}
```

### Visual Question Answering
Ask a question about the image
```json
{
  "model": "salesforce/blip",
  "input": {
    "image": "https://example.com/photo.jpg",
    "task": "visual_question_answering",
    "question": "How many people are in the picture?"
  }
}
```

## Response Format
```json
{
  "job_id": "abc123",
  "status": "pending | processing | completed | failed",
  "result": "URL or data (when completed)"
}
```

## Usage Flow
1. POST /predictions with model and input -> receive job_id
2. GET /predictions/{job_id} -> poll until status is completed or failed
3. Result contains output URL(s)

## Tips
- Captions are short English sentences — great for alt-text and tags
- For longer free-form answers, moondream2 gives richer descriptions
- image_text_matching scores how well a caption fits — useful for ranking

## Documentation
https://replicate.com/salesforce/blip