AccuralAI Documentation

Welcome to the AccuralAI documentation! This documentation covers all packages in the AccuralAI ecosystem.

Quick Start

AccuralAI ships pragmatic tooling that compounds small wins into observable LLM quality improvements. Every package in this monorepo is a building block for evaluating, routing, or operationalizing existing models rather than training new ones.

Packages Overview

Core Packages

Core Packages

Backend Adapters

Backend Adapters

Plugins & Extensions

Plugins & Extensions

Installation

Install the complete AccuralAI ecosystem:

pip install accuralai

Or install individual packages:

pip install accuralai-core
pip install accuralai-cache
pip install accuralai-canonicalize
# ... etc

Philosophy

AccuralAI focuses on:

  • Pragmatic tooling that compounds small wins

  • Observable improvements in LLM quality

  • Building blocks for evaluation, routing, and operationalization

  • Research alongside production code

Getting Started

  1. Core Orchestration: Start with accuralai-core to understand the pipeline architecture

  2. Caching: Add accuralai-cache for production-ready caching

  3. Backends: Choose your LLM provider with accuralai-google or accuralai-ollama

  4. Routing: Implement intelligent routing with accuralai-router

  5. Validation: Add content validation with accuralai-validator

Configuration

See the configuration example for a complete setup.

Contributing

Research and docs live alongside production code so contributors can trace ideas from plan/ into shipped modules.

License

Apache-2.0 License - see individual package licenses for details.