VerSkyLicense inquiry
For Engineers

Build with VerSky.

If you're prototyping a fleet, a flight controller, an air-traffic simulator, or a research tool — here's how the protocol meets you today, and what's already running behind closed doors.

MVP · Pre-release
§1 · Today

Where to start.

The protocol is fully described in published material — enough to start evaluating, prototyping against your own stack, or citing in a paper. The Python MVP is built but not yet publicly distributed (see §2 below).

Read

Whitepaper v1.0

A 23-page protocol description: seven principles, hex topology, AACP, DFR, comparison with classical ATC.

Open the whitepaper →
Follow

Protocol deep-dives

Three technical essays — altitude as information, hex-cell capacity, AACP boundaries.

Browse the blog →
Cite

USPTO filings

Two non-provisional applications (19/551,620 + 19/551,624) — pending examination, publishable Aug 2027.

View filing details →
§2 · Status

Python MVP is built. Public release is gated.

A working Python reference implementation exists today — message parser, 4D reservation primitives, deterministic fallback resolver, and a two-vehicle conflict simulator, with a full unit-test suite. Public distribution is on hold pending the licensing decision; early access is available to qualified teams on request.

Python
MVP built · 53 tests passing

Reference parser, reservation primitives, DFR resolver, 2-vehicle simulator

TypeScript
Planned

Browser/Node.js client SDK, integration with web tooling

C / C++
Planned

Embedded SDK for flight controllers (PX4, ArduPilot)

Rust
Planned

High-throughput cloud workers, fleet operator backends

License: TBD — licensing terms for the reference implementation will be published on the Trust & License page ahead of public release.
§3 · Live Code

What it actually looks like.

Snippets below are from the reference implementation that currently sits behind early-access. Test counts are real — request the SDK to run the full suite locally.

src/versky/dfr.pyBuilt · in tests
def resolve_via_dfr(a: NegotiateInit, b: NegotiateInit) -> str:
    """Both vehicles compute the same resolution.
    Returns vehicle_id of the YIELDING vehicle.

    Hierarchy:
      1. Mission priority class
      2. Remaining energy level
      3. Current heading value
      4. Vehicle identifier (lexicographic)
    """
    rank_a = PRIORITY_RANK[a.mission_priority]
    rank_b = PRIORITY_RANK[b.mission_priority]
    if rank_a != rank_b:
        return b.vehicle_id if rank_a < rank_b else a.vehicle_id

    if a.energy_remaining_pct != b.energy_remaining_pct:
        return (a.vehicle_id
                if a.energy_remaining_pct > b.energy_remaining_pct
                else b.vehicle_id)

    if a.current_heading_deg != b.current_heading_deg:
        return (a.vehicle_id
                if a.current_heading_deg < b.current_heading_deg
                else b.vehicle_id)

    if a.vehicle_id != b.vehicle_id:
        return a.vehicle_id if a.vehicle_id < b.vehicle_id else b.vehicle_id

    return a.vehicle_id  # degenerate case — same identity
Four-level deterministic fallback. Same inputs → same vehicle yields, computed independently on both vehicles (P2 ¶[0020], §5.8).
§4 · Capabilities

What the MVP covers today.

Built capabilities reflect the Python MVP in its current form. Specifications and scope subject to change pending USPTO examination.

Built

Message parsing

AACP Layer 1 (Intent & State Broadcasting) and Layer 2 (Peer-to-Peer Negotiation) message schemas, pydantic-validated.

Built

4D reservation primitives

Hex-cell × floor × direction × time-slice reservations, with constant-cost cell-occupancy lookup and capacity enforcement.

Built

Deterministic fallback resolver

Four-level DFR hierarchy (mission priority → energy → heading → vehicle ID), determinism-tested for symmetric resolution.

Built

Two-vehicle simulator

End-to-end scenario: conflict detection → Layer 1 broadcast → Layer 2 negotiation → DFR fallback → resolution.

Built

Hex grid + 6 directions

Sky-floor mapping (0–150m AGL, six classes), six 60°-separated direction bands with tolerance, per-floor sub-layer math.

Partial

Compliance level switch

Five compliance levels (Basic / Standard / Advanced / Trusted / Full) modelled in schemas. Layers 3–5 not in scope of the MVP and ship in a later phase.

Planned

Multi-vehicle simulator (N>2)

Scale beyond two-vehicle scenarios, asymmetric traffic patterns, capacity-saturated cells. Planned for a later phase.

Planned

Trust scoring + audit log

AACP Layer 4 (Interoperability & Trust) and Layer 5 (Explainability & Audit) — hash-chained decision logging. Planned for a later phase.

§5 · Early Access

Get hands on the MVP.

The Python MVP is not yet on a public repository. Qualified teams — researchers, fleet operators, eVTOL programmes, regulators — can request pre-release access. Tell us what you're building, your fleet size or research scope, and your integration target. Priority goes to use cases whose feedback shapes the public release.