nexmind 0.1.0

NexMind AI — Clean ML training data with AI-powered agents

NexMind AI — DataClean Swarm SDK

Clean your ML training data with AI-powered agents.

Install

pip install nexmind

Quick Start

from nexmind import NexMindClient
import pandas as pd

# Initialize
client = NexMindClient(api_key="dcs_your_key")

# Clean a DataFrame
df = pd.read_csv("dirty_data.csv")
clean_df, report = client.clean(df)

print(f"Quality: {report['quality_score_before']:.0%} -> {report['quality_score_after']:.0%}")
print(f"Rows removed: {report['rows_removed']}")
print(f"Rows repaired: {report['rows_repaired']}")

Choose Agents

# Run specific agents only
clean_df, report = client.clean(
    df,
    agents=["schema", "dedup", "repair"],  # skip anomaly + labeling
)

Async Jobs

# Submit and don't wait
job = client.submit_job(
    name="Nightly data clean",
    source={"type": "s3", "bucket": "raw", "key": "data.csv"},
    output={"type": "s3", "bucket": "clean", "key": "clean.parquet"},
)

# Check later
job.refresh()
print(job.status)  # "running" | "completed" | "failed"

# Or wait
job.wait()
print(job.report)

5 AI Agents

Agent	What it does
schema	Fix column types, names, formats
anomaly	Detect outliers (Z-score, IQR, Isolation Forest)
dedup	Remove exact + fuzzy duplicates
repair	Fill missing values, fix typos, standardize
labeling	Detect PII, classify columns

Get Your API Key

1. Sign up at app.getnexmind.com
2. Go to API Keys page
3. Create a key

Links

• Dashboard
• API Docs
• Website