Quickstart¶
Getting access¶
To use the Neuralk Cloud API, you need an API key. Run the following command to get started:
neuralk login
This will display instructions and a link to create your account at:
https://prediction.neuralk-ai.com/register
Your API key will be generated upon registration. It looks like nk_live_xxxxxxxxxxxx.
Installation¶
The Neuralk SDK is available on PyPI and can be installed using pip:
pip install neuralk
Verifying your installation¶
You can verify that the Neuralk package was installed correctly:
python -c "import neuralk; print('Neuralk imported successfully')"
Setting up your API key¶
You can provide your API key in two ways:
Option 1: Environment variable (recommended)¶
Linux/macOS:
export NEURALK_API_KEY=nk_live_your_api_key_here
Windows:
set NEURALK_API_KEY=nk_live_your_api_key_here
Option 2: Pass directly to the classifier¶
from neuralk import NICLClassifier
clf = NICLClassifier(api_key="nk_live_your_api_key_here")
Your first prediction¶
Here’s a simple example to verify everything is working:
import time
from sklearn.datasets import make_classification
from sklearn.model_selection import train_test_split
from sklearn.metrics import accuracy_score
from neuralk import NICLClassifier
# Generate sample data
X, y = make_classification(random_state=0, n_samples=1_000, n_features=10)
X_train, X_test, y_train, y_test = train_test_split(X, y)
# Create and use the classifier
start = time.monotonic()
clf = NICLClassifier() # Uses NEURALK_API_KEY environment variable
clf.fit(X_train, y_train)
predictions = clf.predict(X_test)
elapsed = time.monotonic() - start
print(f"Fit & predict took {elapsed:.1f}s")
print(f"Accuracy: {accuracy_score(y_test, predictions):.2%}")
Save this as quickstart.py and run:
python quickstart.py
You should see the accuracy score printed after a few moments.
On-premise deployment¶
For on-premise deployments, use the host parameter instead of an API key:
from neuralk import NICLClassifier
clf = NICLClassifier(host="http://your-server:8000")
clf.fit(X_train, y_train)
predictions = clf.predict(X_test)
Next steps¶
Explore the Example Gallery for more detailed usage patterns
Learn about Using the model directly for advanced configurations
Discover Built-in selection of the most informative context for optimizing context selection