API

Use the ModelStudio API to integrate your fine-tuned models into applications.

Authentication

ModelStudio uses bearer tokens for API authentication. These tokens are created within ModelStudio itself.

Creating API Keys

Navigate to Organizational Settings → API Keys
Click Create API Key
Give it a descriptive name
Copy and securely store the token

Security: API keys are shown only once. Store them securely and never commit them to version control.

Using API Keys

Include your API key in the Authorization header:

Authorization: Bearer YOUR_API_KEY

Base URL

https://api.modelstudio.app

API Endpoints

Create Generation Job

POST /api/v1/organizations/<YOUR_ORGANIZATION_ID>/generate

Submit a generation job for processing.

Request Parameters:

{
  "model_id": "your_model_id",
  "messages": [[
    {"role": "user", "content": "Your prompt here"}
  ]],
  "temperature": 0.7,
  "max_tokens": 512
}

model_id (required): Your fine-tuned model ID
messages (required): Array of conversation arrays for batch processing
temperature (optional): 0.0-1.0, controls randomness, default 0.7
max_tokens (optional): Maximum response length, default 512

Response:

{
  "job_id": "job_abc123",
  "status": "CREATED",
  "message": "Generation job created successfully"
}

Retrieve Generation Output

GET /api/v1/organizations/<YOUR_ORGANIZATION_ID>/generate/<JOB_ID>

Poll for job completion and retrieve results.

Response:

{
  "job_id": "job_abc123",
  "job_status": "completed",
  "total_tasks": 1,
  "completed_tasks": 1,
  "results": [
    {
      "task_index": 0,
      "status": "completed",
      "result": "Generated text response",
      "error": null
    }
  ]
}

Code Examples

Python

import requests

API_KEY = "your_api_key"
ORG_ID = "your_org_id"
MODEL_ID = "your_model_id"

# Create generation job
response = requests.post(
    f"https://api.modelstudio.app/api/v1/organizations/{ORG_ID}/generate",
    headers={"Authorization": f"Bearer {API_KEY}"},
    json={
        "model_id": MODEL_ID,
        "messages": [[
            {"role": "user", "content": "Hello, how are you?"}
        ]],
        "temperature": 0.7,
        "max_tokens": 512
    }
)

job_id = response.json()["job_id"]

# Get results
result = requests.get(
    f"https://api.modelstudio.app/api/v1/organizations/{ORG_ID}/generate/{job_id}",
    headers={"Authorization": f"Bearer {API_KEY}"}
)

print(result.json())

import requests
import time

API_KEY = "your_api_key"
ORG_ID = "your_org_id"
MODEL_ID = "your_model_id"

def generate_and_wait(messages, max_wait=60):
    # Submit job
    response = requests.post(
        f"https://api.modelstudio.app/api/v1/organizations/{ORG_ID}/generate",
        headers={"Authorization": f"Bearer {API_KEY}"},
        json={
            "model_id": MODEL_ID,
            "messages": messages,
            "temperature": 0.7,
            "max_tokens": 512
        }
    )

    job_id = response.json()["job_id"]

    # Poll for completion
    start_time = time.time()
    while time.time() - start_time < max_wait:
        result = requests.get(
            f"https://api.modelstudio.app/api/v1/organizations/{ORG_ID}/generate/{job_id}",
            headers={"Authorization": f"Bearer {API_KEY}"}
        )

        data = result.json()
        if data["job_status"] == "completed":
            return data["results"]

        time.sleep(2)

    raise TimeoutError("Job did not complete in time")

# Usage
messages = [[{"role": "user", "content": "Hello!"}]]
results = generate_and_wait(messages)
print(results[0]["result"])

JavaScript

const API_KEY = "your_api_key";
const ORG_ID = "your_org_id";
const MODEL_ID = "your_model_id";

// Create generation job
const response = await fetch(
  `https://api.modelstudio.app/api/v1/organizations/${ORG_ID}/generate`,
  {
    method: "POST",
    headers: {
      "Authorization": `Bearer ${API_KEY}`,
      "Content-Type": "application/json"
    },
    body: JSON.stringify({
      model_id: MODEL_ID,
      messages: [[
        { role: "user", content: "Hello, how are you?" }
      ]],
      temperature: 0.7,
      max_tokens: 512
    })
  }
);

const { job_id } = await response.json();

// Poll for results
const pollResults = async (jobId, maxAttempts = 30) => {
  for (let i = 0; i < maxAttempts; i++) {
    const result = await fetch(
      `https://api.modelstudio.app/api/v1/organizations/${ORG_ID}/generate/${jobId}`,
      { headers: { "Authorization": `Bearer ${API_KEY}` } }
    );

    const data = await result.json();
    if (data.job_status === "completed") {
      return data.results;
    }

    await new Promise(resolve => setTimeout(resolve, 2000));
  }
  throw new Error("Timeout");
};

const results = await pollResults(job_id);
console.log(results[0].result);

cURL

# Create generation job
curl -X POST https://api.modelstudio.app/api/v1/organizations/<ORG_ID>/generate \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model_id": "your_model_id",
    "messages": [[
      {"role": "user", "content": "Hello!"}
    ]],
    "temperature": 0.7,
    "max_tokens": 512
  }'

# Get results
curl https://api.modelstudio.app/api/v1/organizations/<ORG_ID>/generate/<JOB_ID> \
  -H "Authorization: Bearer YOUR_API_KEY"

Batch Processing

Process multiple inputs in a single API call:

{
  "model_id": "your_model_id",
  "messages": [
    [{"role": "user", "content": "First input"}],
    [{"role": "user", "content": "Second input"}],
    [{"role": "user", "content": "Third input"}]
  ]
}

Results are returned in the same order.

Multi-Turn Conversations

Include conversation history:

{
  "model_id": "your_model_id",
  "messages": [[
    {"role": "system", "content": "You are a helpful assistant."},
    {"role": "user", "content": "What is AI?"},
    {"role": "assistant", "content": "AI stands for..."},
    {"role": "user", "content": "Tell me more"}
  ]]
}

Vision Models

For vision models, include images as base64:

{
  "model_id": "your_vision_model_id",
  "messages": [[
    {
      "role": "user",
      "content": [
        {"type": "text", "text": "What's in this image?"},
        {
          "type": "image_url",
          "image_url": {
            "url": "data:image/jpeg;base64,/9j/4AAQSkZJRg..."
          }
        }
      ]
    }
  ]]
}

Error Handling

Common HTTP status codes:

Code	Meaning
401	Invalid or missing API key
403	Insufficient permissions
404	Model or job not found
400	Invalid request format
500	Server error

Code Generation in Web UI

ModelStudio provides ready-to-use code examples directly in the web interface:

Navigate to the Models section
Select your fine-tuned model
Click the code icon
Copy the generated code for your preferred language

Best Practices

Security

Store API keys securely (environment variables, secrets management)
Never expose keys in client-side code
Rotate keys periodically

Performance

Use batch processing for multiple inputs
Implement exponential backoff for retries
Monitor job completion times

Error Handling

Implement retry logic
Validate inputs before submission
Log errors for debugging

Next Steps

Test your model in the web interface first
Start with small-scale API testing
Monitor performance and errors
Scale to production volumes

Support: For API issues or questions, contact manufactAI through the platform.

API

On this page