Skip to main content
In early childhood education and enlightenment learning scenarios, utilizing MiniMax’s text-to-image understanding capabilities and MiniMax-Speech-02 speech synthesis technology, children can “snap a picture, and learn.”

​
Solution

Photo Recognition
  • Take a photo of any object with the camera, call upon the text large model’s image understanding capability to quickly identify the object’s category, features, and background knowledge.
Knowledge Explanation and Foreign Language Learning
  • The model converts the recognition result into a short knowledge card suitable for children’s understanding.
  • Simultaneously, it calls MiniMax-Speech-02 hyper-realistic TTS to provide explanations with a warm and friendly voice.
Bilingual/Multilingual Interaction
  • Supports English, Japanese, and other multilingual outputs, helping children learn while playing and enhancing their interest in foreign language listening and speaking.
Fun Q&A
  • The AI voice assistant can initiate interactive questions based on the recognized item, such as “Do you know why giraffes have such long necks?”

​
Business Value

Immersive Learning

Visual recognition + voice interaction makes the learning experience more intuitive and fun.

Multilingual Enlightenment

Naturally integrates bilingual learning scenarios, helping children master foreign language vocabulary and pronunciation early on.

Adaptive Content

Automatically adjusts the difficulty and length of explanations according to different age groups.

Strong Scalability

In addition to everyday objects, it can also recognize animals and plants, vehicles, scientific instruments, and many other types of things.

​
Core API Capabilities

  1. Text Synthesis Interface: Input an image for recognition. POST https://api.minimax.io/v1/chat/completions
  2. Speech Synthesis Interface: Convert the recognized explanation into speech. POST https://api.minimax.io/v1/t2a_v2

​
Usage Example

Image Understanding and Speech Synthesis 
import base64
import requests
import json

group_id = "Please fill in your group_id"
api_key = "Please fill in your_api_key"

# 1. Upload image and recognize
def recognize_image(image_path):
    with open(image_path, "rb") as img_file:
        img_base64 = base64.b64encode(img_file.read()).decode("utf-8")

    url = f"https://api.minimax.io/v1/chat/completions"
    headers = {
        "Authorization": f"Bearer {api_key}",
        "Content-Type": "application/json"
    }
    payload = {
        "model": "MiniMax-Text-01",
        "messages": [
            {
                "role": "system",
                "content": (
                    "You are an intelligent teacher specializing in explaining things to children.\n"
                    "Your task is to identify objects in pictures and explain them in a simple and fun way.\n"
                    "The answer must include both Chinese and English.\n"
                    "Chinese should be gentle and kind; English should be concise and easy to understand.\n"
                    "For example: '这是一个苹果，它是一种甜甜的水果。'\n"
                    "English: 'This is an apple. It is a sweet fruit.'"
                )
            },
            {
                "role": "user",
                "content": f"Please identify and explain this image: [Image base64:{img_base64}]"
            }
        ]
    }

    response = requests.post(url, headers=headers, data=json.dumps(payload))
    result = response.json()
    text_output = result["choices"][0]["message"]["content"]
    return text_output

# 2. Convert recognized explanation to speech
def text_to_speech(text, output_file="output.mp3"):
    url = f"https://minimax.io/v1/t2a_v2"
    payload = json.dumps({
        "model": "speech-2.5-hd-preview",
        "text": text,
        "stream": False,
        "voice_setting": {
            "voice_id": "English_expressive_narrator",
            "speed": 1,
            "vol": 1,
            "pitch": 0,
            "emotion": "happy"
        },
        "audio_setting": {
            "sample_rate": 32000,
            "bitrate": 128000,
            "format": "mp3",
            "channel": 1
        }
    })
    headers = {
        "Authorization": f"Bearer {api_key}",
        "Content-Type": "application/json"
    }
    response = requests.post(url, headers=headers, data=payload)
    parsed_json = response.json()
    audio_value = bytes.fromhex(parsed_json["data"]["audio"])
    with open(output_file, "wb") as f:
        f.write(audio_value)
    print(f"Speech saved as {output_file}")

# ==== Demo Run ====
if __name__ == "__main__":
    explain_text = recognize_image("test.jpg")  # Input image
    print("Recognition and explanation result:\n", explain_text)

    text_to_speech(explain_text, "output.mp3")