Name: MiniGPT-4 API: Image Chat and Description
Brand: magicapi
SKU: api-magicapi-text-from-images-generator-daanelson-minigpt-4
Availability: InStock
Author: API Market

Shashank Agarwal

# 🖼️ Image Chat and Description API (daanelson MiniGPT-4)

## Overview
MiniGPT-4 is a model that generates text in response to an input image and a prompt. Send an `image` and ask for a description, an answer to a question or a piece of writing about the picture.

Decoding settings let you choose between more predictable text (lower `temperature`, more `num_beams`) and more varied text, and cap the output with `max_new_tokens` and `max_length`.

## Features
- **Image plus prompt**: the answer is grounded in the image you send.
- **Beam search**: `num_beams` (1 to 10, default `3`) explores several candidate answers.
- **Sampling controls**: `temperature`, `top_p` and `repetition_penalty`.
- **Length limits**: `max_new_tokens` caps the generated tokens; `max_length` caps prompt plus output.
- **Single text output**: the response is one string.

## Use Cases
1. **Image descriptions**: describe photos for catalogues, search or alt text.
2. **Questions about an image**: answer what is shown, counted or happening in a picture.
3. **Creative writing from photos**: stories, captions and posts based on an image.
4. **Content review**: summarize what an uploaded image contains.

## Visual Examples
### Example Input
Prompt: "This llama's name is Dave. Write me a story about how Dave found his skateboard."

![Input image: a llama](https://blog.api.market/wp-content/uploads/2026/10/minigpt-4-input-api-market.jpg)

Other fields: `num_beams` `5`, `temperature` `1.32`, `top_p` `0.9`, `repetition_penalty` `1`, `max_new_tokens` `3000`, `max_length` `4000`.

### Example Output
A five-paragraph story. The first paragraph:

> Dave the llama was always looking for new ways to have fun. One day, he stumbled upon a skateboard in the park and decided to give it a try. As he started to ride, he felt the wind in his hair and the thrill of movement in his legs. He quickly became addicted to skating and would spend hours each day riding around the park.

## Input Parameters
| Field | Type | Required | Default | Allowed values | Description |
| --- | --- | --- | --- | --- | --- |
| `image` | string (URL) | Yes |  |  | Image to discuss |
| `prompt` | string | Yes |  |  | Prompt about the input image |
| `num_beams` | integer | No | `3` | 1 to 10 | Number of beams for beam search decoding |
| `temperature` | number | No | `1` | 0.01 to 2 | Temperature for generating tokens; lower values give more predictable results |
| `top_p` | number | No | `0.9` | 0 to 1 | Sample from the top p fraction of most likely tokens |
| `repetition_penalty` | number | No | `1` | 0.01 to 5 | Penalty for repeated words; 1 is no penalty, values above 1 discourage repetition, below 1 encourage it |
| `max_new_tokens` | integer | No | `3000` | minimum 1 | Maximum number of new tokens to generate |
| `max_length` | integer | No | `4000` | minimum 1 | Total length of prompt and output in tokens |

## Pricing
- **Pay As You Go (PAYG) pricing**: you only pay for what you use.
- **Price**: $0.12 per API unit.
- **1 API unit = 1 request with up to 77 seconds of processing.** A request that takes longer uses one unit per started block of 77 seconds (for example, 192 seconds of processing uses 3 units, $0.36).
- Long answers and more `num_beams` take longer and can use more units.
- **Example cost**: the example request took 18.4 seconds of processing, which is 1 unit, $0.12.
- **Failed or cancelled predictions are not billed.**
- **GET requests to check prediction status** are free and don't count toward API unit usage.
- **No monthly subscription fee.**

## API Documentation

### Base URL
All API requests go to: `https://prod.api.market/api/v1/magicapi/text-from-images-generator-daanelson-minigpt-4`

### Authentication
Send your api.market API key in the `x-api-market-key` header:
```
x-api-market-key: YOUR_API_MARKET_KEY
```

### Asynchronous Processing
**IMPORTANT**: This API is asynchronous. The flow is:
1. Send a POST request to `/predictions`. A successful request returns **201 Created** with a prediction `id` and a `status` of `starting` or `processing`.
2. Poll GET `/predictions/{id}` with that ID.
3. Keep polling while `status` is `starting` or `processing`.
4. Stop when `status` is `succeeded` (read `output`), `failed` or `canceled` (read `error`).

A polling interval of 1 second and an overall timeout of 600 seconds (10 minutes) work well.

### Endpoints

#### Create a Prediction
```bash
curl -X POST "https://prod.api.market/api/v1/magicapi/text-from-images-generator-daanelson-minigpt-4/predictions" \
  -H "Content-Type: application/json" \
  -H "x-api-market-key: YOUR_API_MARKET_KEY" \
  -d '{
    "version": "e447a8583cffd86ce3b93f9c2cd24f2eae603d99ace6afa94b33a08e94a3cd06",
    "input": {
      "image": "https://blog.api.market/wp-content/uploads/2026/10/minigpt-4-input-api-market.jpg",
      "top_p": 0.9,
      "prompt": "This llama'\''s name is Dave. Write me a story about how Dave found his skateboard.",
      "num_beams": 5,
      "max_length": 4000,
      "temperature": 1.32,
      "max_new_tokens": 3000,
      "repetition_penalty": 1
    }
  }'
```

Response (201 Created):
```json
{
  "id": "0mn329f0w9rne0d1d0wa2sej1g",
  "version": "e447a8583cffd86ce3b93f9c2cd24f2eae603d99ace6afa94b33a08e94a3cd06",
  "status": "starting",
  "created_at": "2026-10-23T10:05:20.681Z"
}
```

#### Check Prediction Status
```bash
curl -X GET "https://prod.api.market/api/v1/magicapi/text-from-images-generator-daanelson-minigpt-4/predictions/{id}" \
  -H "x-api-market-key: YOUR_API_MARKET_KEY"
```

Response when the prediction has finished (200 OK):
```json
{
  "id": "0mn329f0w9rne0d1d0wa2sej1g",
  "status": "succeeded",
  "output": "Dave the llama was always looking for new ways to have fun. One day, he stumbled upon a skateboard in the park and decided to give it a try. As he started to ride, he felt the wind in his hair and the thrill of movement in his legs. He quickly became addicted to skating and would spend hours each day riding around the park. ...",
  "error": null,
  "created_at": "2026-10-23T10:05:20.681Z",
  "started_at": "2026-10-23T10:06:22.333Z",
  "completed_at": "2026-10-23T10:06:40.720Z"
}
```

`output` is a single string with the generated text. The example above is shortened to its first paragraph.

### Status Codes
- **201 Created**: prediction created (POST `/predictions`). This is the normal success response.
- **200 OK**: prediction details returned (GET `/predictions/{id}`).
- **400 Bad Request**: invalid input, for example a missing `image` or `prompt`, or a `num_beams` above 10.
- **401 Unauthorized**: missing or invalid API key.
- **404 Not Found**: no prediction with that ID.
- **429 Too Many Requests**: rate limit exceeded.
- **500 Internal Server Error** or **503 Service Unavailable**: temporary server problem; retry later.

## Rate Limits
All plans allow 10 requests per second. Requests above that return `429 Too Many Requests`; back off before retrying and honour any `Retry-After` header. Polling a prediction with GET is free but still counts toward the per-second limit, so poll every 1 to 5 seconds rather than in a tight loop.

## Error Handling
- Check the HTTP status code of every request and read the response body for the error message.
- A prediction can also end with `status` set to `failed`; the `error` field then explains why. Failed predictions are not billed.
- Retry 429, 500 and 503 responses with exponential backoff. Fix the input before retrying a 400.

## Implementation Guide
### Example Code

#### JavaScript
```javascript
const apiKey = "YOUR_API_MARKET_KEY";
const baseUrl = "https://prod.api.market/api/v1/magicapi/text-from-images-generator-daanelson-minigpt-4";

async function createAndWaitForPrediction(input) {
  const response = await fetch(`${baseUrl}/predictions`, {
    method: "POST",
    headers: {
      "Content-Type": "application/json",
      "x-api-market-key": apiKey
    },
    body: JSON.stringify({ version: "e447a8583cffd86ce3b93f9c2cd24f2eae603d99ace6afa94b33a08e94a3cd06", input })
  });

  if (response.status !== 201) {
    throw new Error(`API error: ${response.status} ${await response.text()}`);
  }

  const prediction = await response.json();
  return pollForResults(prediction.id);
}

const delay = (ms) => new Promise((resolve) => setTimeout(resolve, ms));

async function pollForResults(predictionId) {
  const maxTimeout = 600000;
  const pollInterval = 1000;
  const startTime = Date.now();

  while (Date.now() - startTime < maxTimeout) {
    const response = await fetch(`${baseUrl}/predictions/${predictionId}`, {
      headers: { "x-api-market-key": apiKey }
    });

    if (!response.ok) {
      throw new Error(`API error: ${response.status} ${await response.text()}`);
    }

    const prediction = await response.json();

    if (prediction.status === "succeeded") {
      return prediction.output;
    }

    if (prediction.status === "failed" || prediction.status === "canceled") {
      throw new Error(`Prediction ${prediction.status}: ${prediction.error || "Unknown error"}`);
    }

    await delay(pollInterval);
  }

  throw new Error("Prediction timed out after 10 minutes");
}

createAndWaitForPrediction({
  "image": "https://blog.api.market/wp-content/uploads/2026/10/minigpt-4-input-api-market.jpg",
  "top_p": 0.9,
  "prompt": "This llama's name is Dave. Write me a story about how Dave found his skateboard.",
  "num_beams": 5,
  "max_length": 4000,
  "temperature": 1.32,
  "max_new_tokens": 3000,
  "repetition_penalty": 1
}).then((output) => console.log(output));
```

#### Python
```python
import time
import requests

API_KEY = "YOUR_API_MARKET_KEY"
BASE_URL = "https://prod.api.market/api/v1/magicapi/text-from-images-generator-daanelson-minigpt-4"


def create_and_wait_for_prediction(input_data):
    headers = {"Content-Type": "application/json", "x-api-market-key": API_KEY}
    response = requests.post(
        f"{BASE_URL}/predictions",
        headers=headers,
        json={"version": "e447a8583cffd86ce3b93f9c2cd24f2eae603d99ace6afa94b33a08e94a3cd06", "input": input_data},
    )

    if response.status_code != 201:
        raise Exception(f"API error: {response.status_code} {response.text}")

    return poll_for_results(response.json()["id"])


def poll_for_results(prediction_id):
    max_timeout = 600
    poll_interval = 1
    headers = {"x-api-market-key": API_KEY}
    start_time = time.time()

    while time.time() - start_time < max_timeout:
        response = requests.get(f"{BASE_URL}/predictions/{prediction_id}", headers=headers)

        if response.status_code != 200:
            raise Exception(f"API error: {response.status_code} {response.text}")

        prediction = response.json()

        if prediction["status"] == "succeeded":
            return prediction["output"]

        if prediction["status"] in ("failed", "canceled"):
            raise Exception(f"Prediction {prediction['status']}: {prediction.get('error') or 'Unknown error'}")

        time.sleep(poll_interval)

    raise Exception("Prediction timed out after 10 minutes")


output = create_and_wait_for_prediction({
    "image": "https://blog.api.market/wp-content/uploads/2026/10/minigpt-4-input-api-market.jpg",
    "top_p": 0.9,
    "prompt": "This llama's name is Dave. Write me a story about how Dave found his skateboard.",
    "num_beams": 5,
    "max_length": 4000,
    "temperature": 1.32,
    "max_new_tokens": 3000,
    "repetition_penalty": 1
})
print(output)
```

## FAQ
- **How do I get shorter answers?**
  - Lower `max_new_tokens`, or ask for a short answer in `prompt`.
- **How do I make the output more predictable?**
  - Lower `temperature` (for example to 0.5) and keep `num_beams` at 3 or more.
- **What is the difference between `max_new_tokens` and `max_length`?**
  - `max_new_tokens` limits only the generated text; `max_length` limits the prompt and the generated text together.
- **What does one request cost?**
  - One unit ($0.12) covers up to 77 seconds of processing. The example took 18.4 seconds, so it cost $0.12.
- **What if my prediction fails?**
  - Failed predictions are not billed. Read the `error` field, adjust the input and try again.

## Contact Us
If you have any questions, reach out to us through the chat on the bottom right of the page. You can also email our support team at hello@magicapi.com.


MiniGPT-4 generates text from an image and a prompt: descriptions, answers to questions or stories, with beam search and sampling controls.