Name: GLM-OCR API: Document OCR, Tables and LaTeX Formulas
Brand: magicapi
SKU: api-magicapi-multimodal-ocr-api-lucataco-glm-ocr
Availability: InStock
Author: API Market

Shashank Agarwal

# 🔠 Document OCR API (Z.ai GLM-OCR)

## Overview
GLM-OCR is a compact 0.9B multimodal OCR model from Z.ai. Its developer reports a score of 94.62 on the OmniDocBench V1.5 document parsing benchmark, first place overall, and notes that it fits on a single T4 GPU.

Send a document `image` and pick a `task`: `text` (text recognition, the default), `formula` (formulas as LaTeX), `table` (table parsing) or `custom` (your own instruction in `custom_prompt`). The API returns the recognized content as one text string.

## Features
- **Text recognition**: `task` `text` reads the text of a document image, keeping its line structure.
- **Formulas**: `task` `formula` returns formulas as LaTeX.
- **Tables**: `task` `table` parses tables in the image.
- **Custom instructions**: `task` `custom` uses your `custom_prompt`. Its developer describes JSON-schema information extraction as one of the model's modes.
- **Output length**: `max_new_tokens` from 16 to 16384 (default 4096).

## Use Cases
1. **Invoices and receipts**: read line items, totals and tax lines into text for bookkeeping.
2. **Scanned documents**: turn page images into editable text.
3. **Math and science material**: convert formulas in images into LaTeX.
4. **Field extraction**: ask for specific fields with a `custom` prompt.

## Visual Examples
### Example Input
![Example input: an image of a printed supplier invoice with a table of line items](https://blog.api.market/wp-content/uploads/2026/10/glm-ocr-input-v3-api-market.png)

Settings: `task` `text`, `custom_prompt` empty, `max_new_tokens` 4096.

### Example Output
```text
Hillside Garden Supply
42 Orchard Lane, Springfield

INVOICE
No. INV-2026-0418
Date: 14 Oct 2026

Bill to: Maple Street Cafe, 7 Maple Street

Item Qty Unit price Amount
Terracotta pot, 30 cm 12 $8.50 $102.00
Potting soil, 20 L bag 6 $11.25 $67.50
Herb seedling tray 4 $14.00 $56.00
Drip irrigation kit 1 $45.90 $45.90

Subtotal $271.40
Tax (8%) $21.71
Total due $293.11

Payment due within 30 days. Thank you for your business.
```

## Input Parameters
| Field | Type | Required | Default | Allowed values | Description |
| --- | --- | --- | --- | --- | --- |
| `image` | string (URL) | Yes |  |  | Input document image to OCR (jpeg, png, gif or webp) |
| `task` | string | No | `text` | `text`, `formula`, `table`, `custom` | OCR mode: text recognition, formulas as LaTeX, table parsing, or your own instruction in `custom_prompt` |
| `custom_prompt` | string | No | `""` |  | Your own instruction for the model, used with `task` `custom` |
| `max_new_tokens` | integer | No | `4096` | 16 to 16384 | Maximum number of tokens to generate |

## Pricing
- **Pay As You Go (PAYG) pricing**: you only pay for what you use.
- **Price**: $0.0014 per API unit.
- **1 API unit = 1 successful request with up to 5 seconds of processing.** A request that runs longer is billed one more unit for each started 5-second block (for example, 8 seconds of processing is 2 units, $0.0028). Dense pages and long outputs take longer.
- The example request above took 9.3 seconds of processing, which is 2 units, so it cost $0.0028.
- **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/multimodal-ocr-api-lucataco-glm-ocr`

### 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/multimodal-ocr-api-lucataco-glm-ocr/predictions" \
  -H "Content-Type: application/json" \
  -H "x-api-market-key: YOUR_API_MARKET_KEY" \
  -d '{
    "version": "305df8d2d6079c27d03dea1c5e716382005153889bcebb5dbc32d1356cbd4eac",
    "input": {
      "task": "text",
      "image": "https://blog.api.market/wp-content/uploads/2026/10/glm-ocr-input-v3-api-market.png",
      "custom_prompt": "",
      "max_new_tokens": 4096
    }
  }'
```

Response (201 Created):
```json
{
  "id": "cvjv8fvv75rge0d1d9tbxsggb0",
  "version": "305df8d2d6079c27d03dea1c5e716382005153889bcebb5dbc32d1356cbd4eac",
  "status": "starting",
  "created_at": "2026-10-23T19:00:52.380Z"
}
```

#### Check Prediction Status
```bash
curl -X GET "https://prod.api.market/api/v1/magicapi/multimodal-ocr-api-lucataco-glm-ocr/predictions/{id}" \
  -H "x-api-market-key: YOUR_API_MARKET_KEY"
```

Response when the prediction has finished (200 OK):
```json
{
  "id": "cvjv8fvv75rge0d1d9tbxsggb0",
  "status": "succeeded",
  "output": "Hillside Garden Supply\n42 Orchard Lane, Springfield\n\nINVOICE\nNo. INV-2026-0418\nDate: 14 Oct 2026\n\nBill to: Maple Street Cafe, 7 Maple Street\n\nItem Qty Unit price Amount\nTerracotta pot, 30 cm 12 $8.50 $102.00\nPotting soil, 20 L bag 6 $11.25 $67.50\nHerb seedling tray 4 $14.00 $56.00\nDrip irrigation kit 1 $45.90 $45.90\n\nSubtotal $271.40\nTax (8%) $21.71\nTotal due $293.11\n\nPayment due within 30 days. Thank you for your business.",
  "error": null,
  "created_at": "2026-10-23T19:00:52.380Z",
  "started_at": "2026-10-23T19:00:53.380Z",
  "completed_at": "2026-10-23T19:00:56.380Z"
}
```

`output` is a single string with the recognized content, with line breaks as `\n`.

### 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`, a `task` that is not in the list or a `max_new_tokens` outside 16 to 16384.
- **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/multimodal-ocr-api-lucataco-glm-ocr";

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: "305df8d2d6079c27d03dea1c5e716382005153889bcebb5dbc32d1356cbd4eac", 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({
  "task": "text",
  "image": "https://blog.api.market/wp-content/uploads/2026/10/glm-ocr-input-v3-api-market.png",
  "custom_prompt": "",
  "max_new_tokens": 4096
}).then((text) => console.log(text));
```

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

API_KEY = "YOUR_API_MARKET_KEY"
BASE_URL = "https://prod.api.market/api/v1/magicapi/multimodal-ocr-api-lucataco-glm-ocr"


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": "305df8d2d6079c27d03dea1c5e716382005153889bcebb5dbc32d1356cbd4eac", "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")


print(create_and_wait_for_prediction({
    "task": "text",
    "image": "https://blog.api.market/wp-content/uploads/2026/10/glm-ocr-input-v3-api-market.png",
    "custom_prompt": "",
    "max_new_tokens": 4096
}))
```

## FAQ
- **Which `task` should I use?**
  - `text` for ordinary documents, `formula` for math (returned as LaTeX), `table` for tables, and `custom` to send your own instruction in `custom_prompt`.
- **What does the output look like?**
  - One text string with line breaks as `\n`.
- **Can I send several pages at once?**
  - No. The model takes one `image` per request. Send each page as its own request.
- **How is a request billed?**
  - By processing time: $0.0014 covers up to 5 seconds, and each further started 5-second block adds one unit.
- **Is the 201 response from POST an error?**
  - No. 201 Created means the prediction was accepted. Use the returned `id` to poll GET `/predictions/{id}` until `status` is `succeeded`, `failed` or `canceled`.
- **What if my prediction fails?**
  - Failed predictions are not billed. Read the `error` field, check that the image URL is reachable 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.


Extract text, formulas (LaTeX) and tables from document images, or run a custom prompt, with Z.ai GLM-OCR. $0.0014 per request up to 5 seconds of processing.