OPEN SOURCE / BUILT IN PUBLIC
Openly licensed software for agent orchestration, GPU compute, explainability and production ML, alongside the community catalogues the Institute maintains. Built in public, governed in the open, used in production.
FIG 01 — AGENT CONTROL PLANE
An open-source control plane for deploying, managing and orchestrating AI agents on Kubernetes, with scoped credentials, sandboxed execution, budget and time-horizon limits, human approval gates and a replayable audit trail.
FIG 02 — GPU COMPUTELINUX FOUNDATION AI & DATA
The general-purpose GPU compute framework for cross-vendor graphics cards, built on Vulkan and hosted by LF AI & Data. It supports asynchronous, mobile-enabled GPU processing.
FIG 03 — ECOSYSTEM CATALOGUE
The community catalogue of production ML and MLOps tooling, curated weekly: 24 categories and more than 550 libraries, from serving and orchestration to explainability and ML security.
FIG 04 — SKILLS MARKETPLACE
An open marketplace of portable agent skills and plugins built on the SKILL.md convention, installing into both Claude Code and Copilot CLI from a single catalogue, with contributions accepted by pull request under automated validation.
FIG 05 — EXPLAINABILITY PIPELINE
An explainability toolbox for machine learning across data analysis, model evaluation and production monitoring, pairing the technical tools with the organisational process around them.
FIG 06 — REGULATION CATALOGUE
An open catalogue mapping the ecosystem of guidelines, principles, codes of ethics, standards and regulation being put in place around artificial intelligence, from national law across fifteen economic areas to practical checklists.
FIG 06 — REGULATION CATALOGUE
An open catalogue mapping the ecosystem of guidelines, principles, codes of ethics, standards and regulation being put in place around artificial intelligence, from national law across fifteen economic areas to practical checklists.
FIG 07 — AGENTIC SYSTEMS CATALOGUE
The community catalogue of open-source tooling for production agentic systems: seven categories covering agentic frameworks, observability, protocols, memory management, security, prompt engineering and interfaces, for teams deploying, monitoring, versioning and scaling agents.
FIG 08 — EDITOR THEME
A Vim and Neovim colorscheme in the visual language of the Institute: the signature teal green on pitch black, and a warm sage paper companion. Full truecolor with a 256-colour fallback, optional transparency, and highlights tuned for Treesitter, LSP and completion popups.
CONTRIBUTE