Model-independent sketch-to-image studies: seed scouting, controlled variation, replayable manifests, and a zero-download local tour.
Production-grade TypeScript AI runtime focused on reliability, governance, and reproducible LLM systems. Multi-provider gateway, agents, RAG, workflows, policy engine, audit trails, and deterministic testing — built for teams shipping AI in production.
A modular, self-contained file format for executable AI prompts, logic, and data.
Deterministic factual substrate for multi-agent AI. Shared evidence-backed facts and reproducible grounding through dataset_hash.
Protected v1.0.0 baseline for ARC-Neuron LLMBuilder — local-first LLM lifecycle tooling for benchmark receipts, candidate promotion, lineage, and governed model improvement.
A reproducible Docker-based pipeline for running machine learning experiments with GPU support. This repository provides pre-configured Docker images, environment files, and scripts for: - Setting up GPU-enabled containers - Running training and inference - Managing environments reproducibly
Holistic Multimodel Domain Analysis: A New Paradigm for Robust, Transparent, And Reliable Exploratory Machine Learning that Considers Cross-Model Variability in Feature Importance Assessment
Git-like branching, deterministic replay, and evidence-driven evaluation for reproducible AI agents.
A moment-selected symbolic field for AI with auditable draws, Dreamspell context, 671 symbolic coordinates, and live Thinking 3/6/9.
Deterministic record and replay for agent model calls, tools, files, approvals, clocks, and random values.
Native llama.cpp orchestration for quantized LLM evaluation, drift analysis, and reproducible model comparison.
Turn any Git repository into a local SWE-bench-style coding-agent benchmark.
Open-source capability foundry for turning complex AI behavior into small, testable, local, and rebuildable systems.
Reliable-AI engineering portfolio spanning agent security, evaluation, replay, energy, climate, and product systems.
To associate your repository with the reproducible-ai topic, visit your repo's landing page and select "manage topics."