MyKAI Node 0.4.6 — Steadier, and honest about what it sees
A reliability-and-honesty release. Your node now rides out VPN reconnects and network blips without restart loop and the app stops crying wolf — antivirus warnings, reachability timing, and status messages now reflect what's really happening. Plus two genuinely new conveniences: auto-pause if your VPN drops, and one-click disk cleanup for networks you're not using.
(Coming from 0.4.4? This rolls up everything in both 0.4.5 and 0.4.6.)
What's new
No more restart loops — rock-solid through VPN changes
Previously, if your VPN reconnected or your IP changed at just the wrong moment, the node could trip over itself and get stuck in a restart loop — two copies racing for the same database file. That's fixed at the root. Restarts now go through a single queue: they happen one at a time, never overlap, collapse together when several things fire at once, and wait until your connection has settled and your node is synced before kicking in. The result is a node that stays up through VPN exit-server switches, brief network drops, and address changes. (Stress-tested with rapid back-to-back restarts — zero lock-ups.)
Realistic timing — inbound peers take 1–3 days, and the app now says so
The honest expectation for a fresh public address is that long-lived incoming peers settle in over 1–3 days (roughly 12–72 hours) — not "a few hours," which a few screens used to imply. The reachability text now says this plainly, and reminds you that changing your address restarts that clock. Fewer nasty surprises, less "why aren't peers sticking yet?"
New: pause your node automatically if your VPN drops
For privacy-conscious operators running behind a VPN, there's a new option in Settings → General — "Pause node if VPN drops (privacy)." With it on, the moment your tunnel goes down the node stops completely, so your real IP never reaches the Kaspa network — and it resumes automatically the instant the VPN reconnects. It's off by default (the default keeps your node running through a drop for maximum uptime); turn it on if hiding your home IP matters more to you than a little downtime. Either way, when a tunnel drops the app now says "VPN tunnel down — inbound paused until it reconnects; your node keeps running," instead of pointlessly trying to restart into a dead tunnel.
The antivirus warning stops crying wolf
The "your antivirus may be slowing your node" banner used to fire on a simple timer — if first sync took more than ten minutes it blamed antivirus, even when sync was progressing fine. It now only appears when there's real evidence of a problem, and it names the actual cause — a slow disk, low memory, or several antivirus products scanning at once — instead of always pointing at antivirus. And it always says to exclude the node, never to disable your protection. (Setup now also flags these up front: a slow data drive like an external or mechanical HDD, low free RAM, or multiple real-time scanners running together — the quiet reasons a first sync can crawl.)
Free up disk space without switching networks
You can now delete the chain data for a network you're not actively running — say, clearing old testnet data while your mainnet node keeps running — straight from settings, without switching networks first and without interrupting or restarting your live node. Each network with stored data is listed with its size and a two-click Delete.
The app tells the truth about what it's doing
A whole set of "stop the UI from lying" fixes:
- If the node goes quiet during sync for more than a minute, the status degrades to a clear "no progress — it may have stalled or lost connection" warning, instead of calmly claiming everything's fine.
- After an unclean shutdown, the index rebuild gets its own honest message ("rebuilding the index — usually done within 30 minutes") instead of borrowing the multi-hour first-sync estimate.
- The diagnostic now includes a restart history (the last several restarts, each with its reason and time), is honest about how your address is advertised, no longer counts automatic VPN restarts as ones you triggered, and stops mislabeling a passive Windows Defender as "missing protection" when a third-party antivirus is in charge.
- Reachability advice is now context-aware: if you're on a VPN, it won't tell you to forward a port on your router (which wouldn't help).
Smoother and quieter — on every language of Windows
- The "Fix clock" button now behaves correctly when you cancel the Windows permission prompt, and works on non-English-language Windows.
- Background system checks no longer briefly flash black command-prompt windows on your screen.
- Fixed a repeated Windows permission (UAC) prompt that could keep reappearing on non-English Windows — it came from a firewall check that couldn't read translated system text, now handled language-independently.
- Added another antivirus setup guidance.
- If your KasMap link ever drops, the app now tells you to re-paste your token to reconnect, instead of quietly disappearing from the map. (Your node keeps running fine; only its map listing pauses.)
Updated Kaspa node software — fork-ready
The bundled Kaspa node software is updated to v2.0.1 (Toccata), keeping you current and ready for the network's upcoming June 30 hard fork. If you're on an older bundled node, updating now keeps you from falling behind the upgraded network.
Hardened end to end
This release also includes a large round of behind-the-scenes hardening from a thorough security-and-stability audit — tighter input validation, safer handling of external system processes, and defensive checks throughout. Nothing you'll click; everything you rely on is sturdier.
Auto-update behavior
If you're on any 0.3.x or 0.4.x version, MyKAI will detect this update on next launch and prompt you to restart. Your data, settings, account key, and node identity are all preserved — and your synced chain data carries over, so there's no full re-sync.
If auto-update doesn't pick it up, download the installer below and run it over your existing install.