Packages

Hexagonal architecture core for building flexible RAG systems in Elixir.

Elixir implementation of the GEPA (Genetic-Pareto) optimizer that combines LLM-powered reflection with Pareto search to evolve text-based system components.

Explainable AI (XAI) tools for the Crucible framework. Includes LIME implementations, SHAP-like explanations, feature attribution, and model interpretability for local and global explanations.

A 100% faithful port of the DuckDB Python client to Elixir, using the DuckDB CLI for full API compatibility with the official Python client.

Multi-model ensemble prediction with voting strategies for AI reliability. Leverages BEAM parallelism for massively concurrent LLM queries.

Experimental research framework for running AI benchmarks at scale. Provides orchestration, streaming processing with Flow/GenStage, and statistical analysis.

Adversarial testing and robustness framework for AI models with 25 attacks (character/word/semantic perturbations, prompt injection, jailbreak, extraction, inversion), defenses (detection/filtering/sanitization), certified robustness metrics, and attack composition.

Request hedging for tail latency reduction in distributed systems. Implements Google's 'Tail at Scale' with adaptive strategies. Reduces P99 latency by 75-96%.

Structured causal reasoning chain logging for LLMs. Captures decision-making processes with cryptographic verification for transparency and debugging.

A library to make building performant RAG systems in Elixir easy

Configuration management with CLI parsing for Elixir

Advanced telemetry collection and analysis for AI research

Model versioning, artifact storage, and lineage tracking for ML pipelines

The command and control layer for the AI-powered enterprise.

Production-ready data validation and quality library for Elixir ML pipelines. Provides 22 built-in expectations, drift detection, advanced profiling with outliers and correlations, statistical analysis, and comprehensive quality metrics for machine learning workflows.

vLLM for Elixir via SnakeBridge - Easy, fast, and cheap LLM serving for everyone. High-throughput LLM inference with PagedAttention, continuous batching, and OpenAI-compatible API.

Application layer for Portfolio ecosystem providing RAG workflows, graph analysis, domain registries, and CLI tools built on portfolio_core and portfolio_index.

Composable regularization penalties and loss functions for the Nx ecosystem

Synapse integration for altar_ai - SDK-backed LLM providers for multi-agent workflows. Provides unified adapter layer for Gemini, Claude, and Codex with automatic fallback, workflow actions, signal handlers, and telemetry bridging.

Model deployment orchestration with health checking and rollback