Skip to content

Audit evidence and explanations you don't have to translate yourself

WhiteBoxXAI packages what your AI actually did (SHAP, LIME and counterfactual explanations, bias audits, drift history, LLM audit trails, governance decisions) into evidence you can hand to an auditor or a board, with every number traceable to the data that produced it. It's also the monitoring platform your engineering team uses to produce that evidence in the first place.

Get started Generate a report Install the SDK

Start here

  • Getting Started

    Create an account, run the readiness setup, connect a model, and generate your first report.

  • The Enhanced Dashboard

    Overview, Needs Attention, Model Register, Assurance, Observability, Compliance & Evidence, Reviews & Decisions and the Risk Register: what each shows and how to use it.

  • Reports

    Board packs, ISO/IEC 42001 Statements of Applicability, EU AI Act Annex IV documentation, model cards and evidence packs, signed off and shared with auditors through verified links.

  • Python SDK

    Register models, log predictions and LLM calls from any provider, and group runs into Governance Events in a few lines.

  • MCP & AI agents

    Give Claude, Copilot, Cursor or your own agent 197 tools over your WhiteBoxXAI data, with agent skills that keep it honest.

  • Tutorials

    Explore with sample data, take a model to an evidence pack, govern an LLM app, or connect an assistant over MCP.

What you can do with WhiteBoxXAI

  • Explain every decision

    SHAP, LIME and DiCE counterfactuals computed from your model, so you can answer why a prediction was made and what would have changed it.

  • Audit for fairness

    Fairness audits across the groups you define, with per-group comparisons and the evidence regulators expect, not a summary of one.

  • Show compliance readiness

    Control catalogs for ISO/IEC 42001, the EU AI Act, GDPR, CCPA and NIST AI RMF, with your applicability decisions and approved evidence counted control by control. Compliance frameworks

  • Govern with confidence

    Review boards with votes signed by an authenticator code, separation of duties, and an immutable decision archive.

  • Track AI risk

    An AI Risk Register with owners, likelihood × impact scoring and a full audit trail, and a Trust Score per model a board can read.

  • Monitor LLMs, agents and models

    Drift, performance, LLM and RAG audit trails, safety scans, prompt versions and multi-agent workflows, with alert rules evaluated as new evidence arrives.

Install the SDK

pip install whitebox-xai-sdk

import os

from whiteboxxai import WhiteBoxXAI

client = WhiteBoxXAI(api_key=os.environ["WHITEBOX_API_KEY"])

Then follow Getting Started to register a model and log your first predictions. AI agents can start from llms.txt.

Need help?