Skip to content
  • MangoFly

    A self-hosted WireGuard mesh. Devices connect straight to each other; the coordination server is one binary and a SQLite file, and never sees their traffic.

    encrypted WireGuard · peer to peerLaptopbehind home NATServerin a datacentrePhoneon mobile datacoordination serverone binary · one SQLite filecontrol plane only (TLS)keys · tunnel addresses · peer lists · sealed ICE candidatesholds no private keys · carries no traffic · cannot decryptdatacontrol
  • MangoDock

    Docker management with nothing on the hosts. Reaches each daemon over an ordinary SSH session — no agent to install, no port to open.

    The MangoDock dashboard showing three host cards with container state counts, CPU and memory gauges, a usage history and recent events
  • MangoWiFi

    A Wi-Fi 6/7/8 test bench. One binary runs as Console or Agent either side of the access point under test, measuring latency under real load.

    AP under testWi-Fi 6 / 6E / 7Agentstation side · real radioLAN receiveriperf3 -sConsoleUI · orchestrates · probes
  • Blog
  • Nothing phones home

    No telemetry, no analytics, no crash reporter, no account login. Check it with a packet capture on your own network.

    Download MangoSSH
  • Project
  • Download
  • Wi-Fi 8 (802.11bn D2.0)

    The eighteen-row catalog, the three live runners, and what UHR is actually asking for.

    The Wi-Fi 8 view keeps its own score: 18 rows, 3 runners wired, and a banner saying not to read a pass as certification-grade.

    The Wi-Fi 8 section tracks draft 802.11bn D2.0, dated July 2026. Its shape is reliability rather than raw speed, which is why the rows that execute today are all latency KPIs rather than throughput records.

    Five categories

    • MAPC — §35.14 — Multi-AP Coordination: coordinated beamforming, coordinated spatial reuse, coordinated TDMA and non-primary channel access. Every row needs a second AP.
    • Enhanced MLO — §35.3 — Adding and removing links on a live association, recovering from a link failure, and binding links across two coordinated APs.
    • Seamless Roaming — §35.11 — State carryover, the MPDU-loss boundary at handoff, re-auth latency and sticky-client prevention.
    • Reliability — WFA UHR KPIs — The p95 floor, endurance with drift, the p99.9 tail, and throughput uplift at range.
    • Power — §35.9 — Enhanced TWT negotiation and low-latency doze.

    What runs today

    • p95 latency floor — 200 pps for 30 s with concurrent load fired once. Passes at p95 ≤ 5 ms, warns to 10 ms; loss passes at 0.1%, warns to 1%. The target comes from the WFA UHR AI/XR figure.
    • 30-minute endurance with drift — Same thresholds over 30 minutes with the load cycled continuously, split into six 5-minute buckets. A run whose later buckets degrade is reported as drifted and warns.
    • p99.9 tail — 30 minutes scored on p99.9: passes at 10 ms, warns to 20 ms. p95 describes the typical bad frame; p99.9 describes the one that makes a headset stutter.

    Read the sample count before quoting the tail

    At 200 packets per second for 30 minutes a run collects 360,000 samples, which puts roughly 360 in the top 0.1%. That is enough to be directional and not enough to certify — the project's own note says about a million would be wanted for that. The figure is reported either way; the interpretation is yours.

    If no LAN receiver is configured the concurrent load skips and the probe still runs. The result message says so explicitly, precisely so an unloaded pass is never read as a UHR compliance signal.