Question 1

What is NeuroBridge and why does it exist?

Accepted Answer

NeuroBridge is an open protocol stack that translates Vision-Language-Action (VLA) model outputs into Spiking Neural Network (SNN) spike trains. It exists because no industry standard bridges the gap between large-scale AI decision-making and neuromorphic hardware for Physical AI applications like robotics.

Question 2

How does VLA→SNN pulse encoding work?

Accepted Answer

NeuroBridge uses a novel pulse encoding algorithm that converts VLA action tokens into temporally-coded spike trains. The protocol uses MsgPack binary encoding to achieve sub-millisecond translation latency while preserving semantic fidelity — the spike patterns maintain the high-level intent of the VLA output.

Question 3

Is NeuroBridge compliant with EU AI Act?

Accepted Answer

Yes. NeuroBridge is designed for EU AI Act Art.50 (GPAI transparency, effective 2026-08-02) and Art.14 (high-risk conformity assessment) compliance. We are signing the GPAI Code of Practice by the 2026-07-22 deadline. The Agent Governance Layer provides automated compliance logging and transparency reporting.

Question 4

What is the Agent Governance Layer (AGL)?

Accepted Answer

AGL is NeuroBridge's on-chain trust engine for AI agents operating in the physical world. It provides: identity verification, compliance status tracking, trust scoring, and legal-policy bridging. It maps regulatory requirements (EU AI Act, ISO standards) directly to executable code constraints.

Question 5

Is NeuroBridge participating in IETF standardization?

Accepted Answer

Yes. NeuroBridge is actively participating in IETF discussions around AI agent protocols, specifically the AI Agent Capabilities (AIC) and SD-JWT for AI identity. We have published strategic analysis and comment drafts for IETF consideration.

Question 6

How does NeuroBridge handle dual-model safety (VLA vs SNN)?

Accepted Answer

NeuroBridge defines formal responsibility boundaries between VLA (decision layer) and SNN (execution layer). When a safety incident occurs, the system can deterministically attribute whether the fault originated in the VLA's decision or the SNN's execution. This enables ISO 13849 / IEC 61508 safety certification.

Question 7

What robots and hardware does NeuroBridge support?

Accepted Answer

NeuroBridge is hardware-agnostic at the protocol level. The SDK supports integration with any VLA model (RT-2, Octo, OpenVLA, π0) and any SNN framework (Norse, SpikingJelly, Intel Lava). Target hardware includes neuromorphic chips (Intel Loihi, IBM NorthPole) and traditional GPUs.

Question 8

How does NeuroBridge relate to BDE Score and IPO Compliance?

Accepted Answer

NeuroBridge is the protocol infrastructure layer. BDE Score™ uses NeuroBridge's compliance framework for its EU AI Act Art.50 transparency. IPO Compliance Service uses NeuroBridge's 16-dimension diagnostic framework for Physical AI pre-IPO assessments. Together they form a compliance triangle: NeuroBridge=protocol, IPO=methodology, BDE=quantitative validation.

Question 9

Is the protocol open source?

Accepted Answer

The NeuroBridge protocol specification is openly published. The SDK is available under a partnership agreement for integration. The core Python SDK, test suite, and documentation are available on GitHub.

Question 10

What is the Legal-Policy Bridge (LPB)?

Accepted Answer

LPB is a component that translates legal and regulatory requirements into machine-executable constraints. It uses Cedar-style policies to encode EU AI Act articles, ISO standards, and industry regulations, allowing automated compliance checking at runtime without legal expertise.

NeuroBridge FAQ

What is NeuroBridge and why does it exist?

How does VLA→SNN pulse encoding work?

Is NeuroBridge compliant with EU AI Act?

What is the Agent Governance Layer (AGL)?

Is NeuroBridge participating in IETF?

How does dual-model safety work?