Skip to main content
This guide takes you from zero to your first prediction in ~5 minutes.

​
1. Get an API key

Sign in to the console, open API Keys → Create, and copy the key (hk_live_...). It is shown once. Store it somewhere safe.

​
2. Install the SDK

Hollerith is a private SDK. It is not on PyPI. During the closed beta it ships as a pre-built wheel served from your control plane (no repo access needed). Your console’s Quickstart tab shows the exact, current URL. It looks like: 
pip install https://hollerith.monarcha.ai/sdk/hollerith-<version>-py3-none-any.whl

​
3. Configure

The client reads two environment variables (or pass them as constructor args): 
export HOLLERITH_API_KEY="hk_live_…"                        # from step 1
export HOLLERITH_BASE_URL="https://hollerith.monarcha.ai"   # prod API origin
base_url is required. There is no public default origin yet. Your console’s Quickstart tab shows the exact value for your environment.

​
4. Your first prediction

Any labeled CSV works. Using the bundled Iris sample: 
from hollerith import Hollerith
import pandas as pd

clf = Hollerith()                       # reads HOLLERITH_API_KEY + HOLLERITH_BASE_URL
df = pd.read_csv("iris.csv")

clf.fit(df, target="species")           # no training step
preds = clf.predict(df.drop(columns="species"))
print(preds[:5])                        # -> ['setosa' 'setosa' 'versicolor' ...]
predict blocks while the job runs and returns predictions row-aligned with your input. The first call after the service has been idle may take longer while the worker warms up. The SDK waits through it automatically.

​
5. Check accuracy (optional)

clf.fit(df, target="species", evaluate=True)
print(clf.evaluation_)                  # held-out / k-fold metric
Your data is ephemeral. The training table and rows you send are purged as soon as the job finishes; only hashes and counts are retained for metering.