Build with Fount

Step 1: Create your account

Go to the Fount Developer Portal and click Sign up.

Verify your email using the OTP sent to you and complete your basic account setup.

Step 2: Create a project

After login, create your first project.

1. Goto "Manage Projects" and click "Create" on the right hand side
2. Fill in the project details:
   • Project Name: A unique identifier for your project
   • Description: A brief description of the project purpose
3. Click "Confirm"

Projects help you organize your datasets, models, and API keys in one place.

Step 3: Generate your API key

1. Navigate to API Keys inside your project
2. Click "Create new Secret Key"
3. Enter a name for your key (e.g., "Development Testing") — the Project field will be auto-filled
4. Click "Confirm"
5. Copy and save your key immediately, it won't be shown again

Step 4: Install the SDK

Install the Python SDK to start building:

pip install fount-core

Step 5: Set up authentication

Choose the appropriate method for your environment:

export FOUNT_API_KEY="your_api_key_here"

Step 6: Test your connection

Verify your setup is working before training a model:

from fount import Fount
client =Fount()
try:
  datasets=client.get_all_datasets()
  print("Connection successful")
  print("Datasets in project :{len(datasets)}")
except Exception as e:
  print("Connection failed. Check FOUNT_API_KEY and network")
  print("Error : {e}")
raise

If this prints successfully, you're ready to start building.

Step 7: Train your first model

Here's a complete example that uploads a dataset and trains a model:

import pandas as pd
from fount import Fount
import time

# Initialize client (uses FOUNT_API_KEY from environment)
client = Fount()

# Upload a dataset
df = pd.read_csv("my_data.csv")
dataset = client.upload_dataframe(df, name="My Dataset")

# Train a model
job = client.train(
    dataset=dataset,
    series_id_cols=["ProductCategory", "Region"],
    categorical_cols=["ProductCategory", "Region","Year","Month"],
    model_name="Q4_sales",
    date_column="Date",
    target="Sales",
    validation_data_required=True,
    validation_split=0.2,
)

# Monitor training
while job.status().get("state") not in ["completed", "failed"]:
    print("Training in progress:", job.status().get("state"))
    time.sleep(30)

if job.status().get("state") == "completed":
    metrics = job.metrics()
    print("Training completed:", metrics)
else:
    print("Training failed:", job.status())

Training runs asynchronously. The loop above polls the job status every 30 seconds until the job completes or fails.

Troubleshooting

SDK Installation

ModuleNotFoundError: No module named 'fount'

The fount-core package is not installed in the Python environment your notebook or script is using.

Fix: Install it in the same environment that runs your code:

pip install fount-core

If you are in a Jupyter notebook, restart the kernel after installing. Verify with:

pip show fount-core

Authentication

FOUNT_API_KEY not set / 401 Unauthorized / Invalid API key

The SDK cannot find or validate your API key.

Common causes:

• Environment variable not set before initializing the client
• Key copied with leading/trailing spaces
• Key belongs to a different project or has expired

Fix: Set the key in the same shell session or notebook cell where you call Fount():

import os
os.environ["FOUNT_API_KEY"] = "your_api_key_here"  # no extra spaces

Then verify by running os.environ.get("FOUNT_API_KEY"), it should return your key, not None.

Next steps

Now that your first model is trained, here's where to go next:

• Explore the API Reference for all available endpoints and parameters
• Learn how to run predictions and inference with your trained model in Fount