fiddler-client 3.12.0rc1

Python client for Fiddler Platform

fiddler-client

The official Python client for Fiddler, the enterprise-grade AI Observability platform. Monitor, analyze, and protect your ML models, LLMs, GenAI applications, and AI Agents in production.

Platform Features

• 🚀 Easy Integration - Simple Python API for model and GenAI application onboarding
• 📊 Real-time Monitoring - Stream production events for drift detection, performance tracking, and anomaly detection
• 🎯 Data Drift & Integrity - Detect distribution changes, data quality issues, and schema violations
• 📈 Custom Metrics - Define and track business KPIs, custom performance metrics, and model-specific measurements
• 🔍 Data Segments - Analyze model behavior across cohorts with custom segments and slices
• 🤖 GenAI & LLM Monitoring - Track hallucinations, toxicity, PII leakage, prompt injection, and response quality
• 🛡️ Guardrails - Real-time protection against unsafe outputs and policy violations in LLM applications
• 📊 Enrichments - Add embeddings, sentiment, topics, and other ML-derived features to your data
• 🔄 Model Comparison - Compare performance across model versions, champion/challenger analysis
• 🎨 Custom Dashboards - Build tailored monitoring views with charts, alerts, and analytics
• 🔔 Smart Alerts - Multi-channel notifications (Slack, Email, PagerDuty) with customizable thresholds
• 🤝 MLOps Integrations - Native support for MLflow, SageMaker, Vertex AI, and other platforms

Installation

pip install fiddler-client

Requirements

• Python 3.10 or higher
• pip 19.0 or higher (latest release preferred)

Quick Start

import fiddler as fdl

# Initialize the client
fdl.init(
    url='https://your-company.fiddler.ai',
    token='your-api-token'
)

# List the existing projects
for project in fdl.Project.list():
    print(project.name)

Documentation

• 📚 Complete Documentation
• 🚀 Getting Started with LLMs
• 🚀 Getting Started with ML
• 🤖 Getting Started with Agentic Monitoring
• 📖 API Reference
• 💡 Example Notebooks

Example Notebooks

Check out our GitHub repository for Jupyter notebooks demonstrating:

• Model onboarding and monitoring setup
• Drift detection and root cause analysis
• Custom metrics and alerting
• LLM and GenAI monitoring
• Integration with popular ML frameworks

Version History

See our release notes for detailed version history.

Support

• 📧 Email: [e-mail]
• 💬 Community: Join our Slack

License

This package is proprietary software. Please refer to your Fiddler license agreement for terms of use.

---

Want to see Fiddler in action? Request a demo