TIFA AI - Online Trial

Developer Documentation

Welcome to the TIFA AI API. This documentation will guide you on how to integrate and use our services. TIFA AI provides powerful natural language processing capabilities, supporting various application scenarios.

Current Version: v1.0 Last Updated: 2024-03-20

Main Features

✨ Supports streaming output, real-time response
✨ Long context support, up to 100K tokens
✨ Multi-language support, including Chinese, English, Japanese, Korean, and 40+ other languages
✨ 99.9% service availability guarantee

Authentication

All API requests require authentication using an API key. You can obtain your API key from the console.

⚠️ Please keep your API key secure, do not store it in plain text in client-side code.

ℹ️ Leftnorth's international business is officially designated to be handled by Swarmeta-ai.me, please use with confidence.

Notes:

API keys have full account permissions, do not disclose them
It is recommended to store API keys in environment variables or configuration files
Regularly rotate API keys to enhance security
You can create different API keys for different applications

Rate Limits

To ensure service quality, API access has the following limits:

Each API key can make up to 60 requests per minute
Maximum input of 4000 tokens per request
Monthly usage limits vary based on the plan

💡 For higher quotas, please contact the business team to upgrade your plan

Best Practices

1. Error Handling

Implement comprehensive error handling mechanisms
Retry network errors
Set reasonable timeout periods
Log detailed error logs

2. Performance Optimization

Use connection pooling to reuse connections
Enable gzip compression
Implement request caching
Adopt asynchronous processing methods

3. Security Recommendations

Store keys in environment variables
Enable HTTPS transmission
Implement request signing mechanisms
Regularly update API keys

4. Development Recommendations

Use SDKs instead of directly calling the API
Isolate development and production environments
Set model parameters appropriately
Implement graceful degradation mechanisms

💡 These best practices will help you build more stable and reliable applications

API Endpoint

POST http://visionsic.com:8001/v1/chat/completions

Request Notes:

Only supports HTTPS protocol
Request body must be in JSON format
Response encoding is UTF-8
Supports cross-origin requests (CORS)

Example Code



import requests

import json



url = "http://visionsic.com:8001/v1/chat/completions"

headers = {

"Content-Type": "application/json",

"Authorization": "Bearer YOUR_API_KEY"

}

data = {

"model": "/home/fangxin/claude",

"messages": [

    {"role": "system", "content": "You are an assistant"},

    {"role": "user", "content": "Hello"}

],

"temperature": 0.4,

"stream": True

}



# Error Handling

try:

    response = requests.post(url, headers=headers, json=data)

    response.raise_for_status()  # Check for HTTP errors

    for line in response.iter_lines():

        if line:

            print(json.loads(line))

except requests.exceptions.RequestException as e:

    print(f"API request error: {e}")

except json.JSONDecodeError as e:

    print(f"JSON parsing error: {e}")



const url = 'http://visionsic.com:8001/v1/chat/completions';

const headers = {

'Content-Type': 'application/json',

'Authorization': 'Bearer YOUR_API_KEY'

};

const data = {

model: '/home/fangxin/claude',

messages: [

    {role: 'system', content: 'You are an assistant'},

    {role: 'user', content: 'Hello'}

],

temperature: 0.4,

stream: true

};



// Using async/await and error handling

async function chatWithAI() {

    try {

        const response = await fetch(url, {

            method: 'POST',

            headers: headers,

            body: JSON.stringify(data)

        });

        

        if (!response.ok) {

            throw new Error(`HTTP error! status: ${response.status}`);

        }

        

        const result = await response.json();

        console.log(result);

    } catch (error) {

        console.error('Request error:', error);

    }

}



chatWithAI();



# Basic Request

curl -X POST http://visionsic.com:8001/v1/chat/completions \

-H "Content-Type: application/json" \

-H "Authorization: Bearer YOUR_API_KEY" \

-d '{

    "model": "/home/fangxin/claude",

    "messages": [

        {"role": "system", "content": "You are an assistant"},

        {"role": "user", "content": "Hello"}

    ],

    "temperature": 0.4,

    "stream": true

}'



# Request with Timeout Settings

curl -X POST http://visionsic.com:8001/v1/chat/completions \

--connect-timeout 10 \

--max-time 30 \

-H "Content-Type: application/json" \

-H "Authorization: Bearer YOUR_API_KEY" \

-d @request.json

Code Notes:

The example includes basic error handling
It is recommended to add retry mechanisms in production environments
Adjust timeout settings as needed
Use asynchronous processing for stream mode

Chat API

Endpoint: http://visionsic.com:8001/v1/chat/completions

Method: POST

Content-Type: application/json

The chat API supports single and multi-turn conversations, allowing control over AI behavior through various parameters. It supports both streaming and regular output modes.

Parameter Explanation

Parameter Name	Type	Required	Description
model	string	Yes	Model path, e.g., "/home/fangxin/claude"
messages	array	Yes	Conversation history, containing role and content fields
temperature	number	No	Temperature parameter, controls randomness, range 0-1, default 0.7
stream	boolean	No	Whether to enable streaming output, default false
max_tokens	integer	No	Maximum length of generated text, default 2048

Response Format:



{

    "id": "chat_12345678",

    "object": "chat.completion",

    "created": 1679478935,

    "model": "/home/fangxin/claude",

    "choices": [{

        "message": {

            "role": "assistant",

            "content": "AI's response content"

        },

        "finish_reason": "stop",

        "index": 0

    }]

}

Error Handling

Common Error Codes:

Error Code	Description	Handling Suggestion
401	Unauthorized	Check if the API key is correct
429	Too Many Requests	Implement request rate limiting
500	Server Error	Retry later or contact support

Advanced Features

1. Streaming Output

Supports real-time streaming of AI responses, providing a better interactive experience.



// Enable streaming output

const response = await fetch(url, {

    method: 'POST',

    headers: headers,

    body: JSON.stringify({...data, stream: true})

});



const reader = response.body.getReader();

while(true) {

    const {done, value} = await reader.read();

    if (done) break;

    console.log(new TextDecoder().decode(value));

}

2. Function Calling

Supports custom function capabilities, expanding the functionality of the AI assistant.



const functions = [{

    "name": "get_weather",

    "description": "Get weather information",

    "parameters": {

        "type": "object",

        "properties": {

            "location": {

                "type": "string",

                "description": "City name"

            }

        }

    }

}];



const response = await fetch(url, {

    method: 'POST',

    headers: headers,

    body: JSON.stringify({

        ...data,

        functions: functions,

        function_call: "auto"

    })

});

3. Context Memory

Supports multi-turn conversation memory, achieving a coherent conversation experience.



const messages = [

    {"role": "system", "content": "You are an assistant"},

    {"role": "user", "content": "How's the weather today?"},

    {"role": "assistant", "content": "It's sunny in Beijing today"},

    {"role": "user", "content": "What about the temperature?"} // Continuing the context

];

4. Model Parameter Tuning

Provides rich parameter configurations for personalized output control.

temperature: Controls output randomness (0-1)
top_p: Controls output diversity
max_tokens: Controls output length
presence_penalty: Controls topic freshness

💡 Advanced features may require higher API permissions, please contact customer service to upgrade