epistemic-ai

Make AI agents and AI workflows measurably reliable. Epistemic measurement, Noetic RAG, Sentinel gating, and grounded calibration for Claude Code and beyond

Free & Open Source AI Agent — Self-Hosted, Self-Learning, No Vendor API. Qwen2.5-7B + LoRA. 35 tools. Built on Islamic Epistemology (IHOS).

The left hemisphere. Frameworks, logic, and certainty architecture. Home of FSVE, AION, LAV, ASL, GENESIS, TOPOS, and 60+ epistemically validated frameworks built to make AI systems reliable, not just capable.

Volute Reasoning Engine (VRE): Structural Epistemic Enforcement for Agentic AI

A local personal memory lab with programmable fractal search and evidence governance.

PassiveQA is a novel question-answering framework that teaches language models when to answer, ask for clarification, or refuse to answer - instead of always attempting to generate responses.

Decoder-only LLM with integrated epistemic tomography. Knows what it doesn't know.

A conceptual framework for epistemically stable AI architectures beyond probabilistic LLM systems. Introducing Stable Epistemic Cognitive Architecture: a layered architecture for reproducible knowledge generation, validation, consolidation, and persistent cognitive reasoning in system-critical/-relevant domains.

The security layer. Every output clears here before it exits. Threat detection, adversarial pattern recognition, red-team archive, and the Go/No-Go authority that can halt the entire system. Nothing bypasses it.

The limbic layer. Personality, register, and internal state monitoring. ALBEDO lives here — the session instrument that governs tone, emotional signal detection, and the felt layer of every response.

The memory layer. Stores the FCL validation archive, routing event history, and the permanent record of everything the brain has learned. What survives here shapes how the whole system routes next time.

An epistemic AI lab for scientifically evaluating how language models track belief, truth, knowledge, inference, defeaters, and luck.

The relay station. Every signal enters here first — classified, routed, and dispatched to the correct brain region before processing begins. The thalamus does not think. It decides where thinking goes.

The refinement layer. Precision after clearance. Every output passes through here for LAV gate validation, density calibration, and roughness preservation — the layer that makes the output worthy of the input.

The presentation layer. Structure, format, and register conversion. The last layer before output — deciding how the brain speaks, not just what it says. Prose or list. Dense or clear. Report or reply.

Epistemic Infrastructure for Distributed Knowledge Production

The official CoDRAG plugin for Paperclip. Give every AI agent deep structural codebase knowledge, blast radius analysis, and semantic search via MCP.

Free & Open Source AI Agent — Self-Hosted, Self-Learning, No Vendor API. Qwen2.5-7B + LoRA. 35 tools. Built on Islamic Epistemology (IHOS).

NoeRelay by Axiovex Systems, LLC — local AI routing, agent execution, and managed coding clients.

Store-neutral protocol for what an agent is entitled to treat as warranted. Five axes — acceptance, conflict, verification, currency, sufficiency — over a bounded evidence view. Memory stores adapt to EWP. EWP does not adapt to them.