verifiable-ai

OML 1.0 via Fingerprinting: Open, Monetizable, and Loyal AI

Offline, on-device AI desktop agent with a signed, independently-verifiable audit trail — runs sealed with zero verified network egress; every action is hash-chained and checkable offline. MIT · pre-launch.

Wisdom before intelligence. A provenance-aware reasoning layer that abstains instead of fabricating — stops LLMs inventing attributions and merging distinct traditions. Measured: 0% fabrication on unknown-answer traps (vs 17–25% raw); −12.5pt hallucinated attributions at 0% false-positive cost. An AGI-candidate proof package, not proven AGI.

The frozen cryptographic kernel of the AXM protocol. Signed, content-addressed knowledge shards with post-quantum signatures (ML-DSA-44) and mechanically enforced non-selective recording. The gold shard is the definition of correctness.

The open standard for AI agent integrity. Evaluate, enforce, and prove that autonomous agents are adversarially coherent, environmentally portable, and verifiably assured.

Run AI agents locally with built-in governance and verifiable proof. Loom + Kernel + Intelligence.

VELLA = Verifiable Evidence Ledger & Logic Authority; deterministic pre-execution authorization and signed proof for AI agents and automated workflows.

Verifiable autonomous vehicle perception using zk-SNARKs (Groth16) with cryptographic integrity and privacy guarantees.

Bundle pack/unpack for portable agent execution: trace + receipts + manifests.

Constitutional substrate for the 5QLN Foundation — sealed Codex anchor, multi-language verifiers, full architecture specification

A Python agent that analyzes crypto markets with AI and proves it didn't make anything up. Deterministic math engine + verifiable SHA-256 audit trail on every call.

Walkthrough demo for the execution-evidence path in the Digital Biosphere Architecture.

Reproducible AVeriTeC + FEVER 1.0 evaluation harness for AgentOracle. If you can't re-run it, it isn't a benchmark.

Third WEB project for yandex-lyceum

An open schema for verifiable financial AI decisions — hash-chained evidence, deterministic reports, no LLM in the fact path. Running on live money.

Deterministic-replay validator for the ERC-8004 ValidationRegistry — settle an agent outcome-claim by replaying a committed function over public data, anchored to Bitcoin. No stake, no committee, no TEE.

Ollmo is a state-native runtime for local AI with Ollama, MLX, and llama.cpp. It makes supported model and media workflows durable and inspectable by recording which outputs exist, what remains open, and why.

Cryptographic receipts for agent traces: hash chain, Ed25519, JCS canonical.

Provider-neutral attribution, source-footer, and royalty settlement system for AI outputs.

The day autonomous agents can produce cryptographically bound, deterministic, machine-checkable deployment certificates is the day CI/CD stops being the authority of deployment and becomes the evidence layer for proof-carrying deployment.