KASPilot

 _  __    _    ____  ____  _ _       _
| |/ /   / \  / ___||  _ \(_) | ___ | |_
| ' /   / _ \ \___ \| |_) | | |/ _ \| __|
| . \  / ___ \ ___) |  __/| | | (_) | |_
|_|\_\/_/   \_\____/|_|   |_|_|\___/ \__|

KASPA OPS TERMINAL  ::  ASIC FLEET CONTROL  ::  GPU SUPERVISOR  ::  CPU DEV MINER

KASPilot is a Kaspa-only mining operations terminal built for ASIC fleet visibility, GPU engine supervision, pool validation, and local kHeavyHash benchmarking. It gives you a production-facing fleet controller, managed GPU lane, and polished CPU dev miner in one Rust binary.

The production lane is ASIC operations. The GPU lane supervises optimized external kHeavyHash engines. The CPU lane is for benchmarking, pool testing, stratum validation, and development work.

UI Preview

Signal

kaspa kheavyhash asic-mining gpu-mining stratum stratum-ssl cgminer-api fleet-monitoring cpu-benchmark autotune rust tui mining-ops

Command Deck

Mission	Command	What it does
Fleet control	kaspa-miner --fleet	Monitor ASIC reachability, TH/s, temp, fan, uptime, pool, accepted/rejected shares
One-shot audit	kaspa-miner --fleet --fleet-once	Poll every configured ASIC once and exit
GPU control	kaspa-miner --gpu	Supervise an installed Kaspa kHeavyHash GPU engine with watchdog restarts
GPU discovery	kaspa-miner --gpu-info	Print local NVIDIA/AMD/macOS GPU runtime hints
CPU benchmark	kaspa-miner --benchmark	Measure local kHeavyHash throughput without touching a pool
CPU autotune	kaspa-miner --tune	Sweep thread and batch settings, then rank the fastest configs
Dev mining	kaspa-miner	Run the Kaspa Common Stratum CPU dev miner with TUI
Plain logs	kaspa-miner --no-tui	Run without dashboard rendering for tmux, services, and logs

What It Has

Layer	Capability
Coin target	Kaspa only
Fleet telemetry	CGMiner-compatible summary, pools, devs, stats over TCP
ASIC inventory	TOML registry with model, location, API/web ports, expected TH/s, enabled state
Pool transport	stratum+tcp://, stratum://, tcp://, stratum+ssl://, ssl://
Pool hardening	Connect timeout, TLS support, TCP_NODELAY, reconnect loop
Hash path	kHeavyHash PoW using kaspa-hashes primitives
GPU lane	External optimized kHeavyHash engine supervisor with config substitution and watchdog restart
Difficulty	Share target derived from mining.set_difficulty
Nonce scan	Extranonce support plus non-overlapping per-thread stride
Operator UI	Ratatui dashboard, plain logs, fleet report, benchmark report
Performance knobs	Release LTO, native CPU build option, configurable batch size, autotune matrix
Releases	Linux archive, signed Windows archives, notarized macOS universal installer

Download

Grab the latest binaries from:

https://github.com/gryszzz/KasPilot/releases/latest

Release assets:

• kaspa-miner-macos-universal.pkg
• kaspa-miner-x86_64-unknown-linux-gnu.tar.gz
• kaspa-miner-x86_64-pc-windows-msvc.zip
• kaspa-miner-aarch64-pc-windows-msvc.zip
• SHA256SUMS.txt

Each release archive includes the binary, README.md, start-mining.toml, config.example.toml, gpu.example.toml, and fleet.example.toml. Windows archives also include a first-run installer that puts the binary on your user path. The macOS package installs starter configs to /usr/local/share/kaspilot/.

First Run

macOS, recommended universal package:

sudo installer -pkg kaspa-miner-macos-universal.pkg -target /
kaspa-miner --version
cp /usr/local/share/kaspilot/start-mining.toml ./config.toml

Windows:

.\install-windows.cmd
kaspa-miner --version

Public desktop releases are blocked unless Apple notarization and Windows Authenticode signing secrets are configured. See docs/macos-no-warning-release.md and docs/windows-no-warning-release.md.

If macOS says Apple could not verify kaspa-miner is free of malware, the file was not notarized or was not the notarized .pkg release asset. Use the signed kaspa-miner-macos-universal.pkg from the latest release after the Apple signing secrets are configured.

Start Mining Config

Release packages include start-mining.toml as the dedicated Kaspa pool and wallet config.

cp start-mining.toml config.toml
$EDITOR config.toml
kaspa-miner --config config.toml

In the macOS installer package, the starter config is installed here:

cp /usr/local/share/kaspilot/start-mining.toml ./config.toml

GPU Mining

KASPilot now has a managed GPU lane for Kaspa. It does not ship a fake GPU kernel; instead it supervises a real installed kHeavyHash GPU engine, injects your Kaspa pool/wallet config, prefixes logs, and restarts the process if it exits.

Create configs:

cp start-mining.toml config.toml
cp gpu.example.toml gpu.toml

Edit config.toml:

pool = "stratum+ssl://YOUR_KASPA_POOL_HOST:5555"
wallet = "kaspa:your_wallet_address"
worker = "kaspa-gpu-rig-01"

Edit gpu.toml to point at the GPU engine you installed:

command = "YOUR_GPU_MINER_BINARY"
devices = "0"
restart = true
restart_delay_secs = 10

args = [
  "--algo", "kheavyhash",
  "--pool", "{pool}",
  "--user", "{login}",
  "--devices", "{devices}",
]

Run it:

kaspa-miner --gpu --config config.toml --gpu-config gpu.toml

Discovery:

kaspa-miner --gpu-info

Placeholders available in gpu.toml: {pool}, {wallet}, {worker}, {login}, {devices}.

Important: Kaspa is ASIC-dominant now. GPU mining is useful for learning, testing, spare hardware, and managed rigs, but serious Kaspa production mining usually means ASICs. KASPilot keeps GPU support honest by treating native CUDA/OpenCL kernels as a dedicated future backend instead of overstating CPU-class code.

Install From Source

git clone https://github.com/gryszzz/KasPilot.git
cd KasPilot
cargo build --release

For best local benchmark numbers on the same machine:

RUSTFLAGS="-C target-cpu=native" cargo build --release

Create local configs:

cp start-mining.toml config.toml
cp gpu.example.toml gpu.toml
cp fleet.example.toml fleet.toml

ASIC Fleet Control

Edit fleet.toml:

poll_secs = 30
timeout_ms = 750

[[devices]]
name = "kas-rig-01"
host = "192.168.1.101"
model = "KS-series"
location = "rack-a"
expected_hashrate_ths = 9.5
api_port = 4028
web_port = 80
enabled = true

Run a single fleet scan:

./target/release/kaspa-miner --fleet --fleet-once

Run the controller continuously:

./target/release/kaspa-miner --fleet

If an ASIC exposes a CGMiner-compatible API on api_port, KASPilot normalizes live TH/s, average TH/s, temperature, fan RPM, uptime, pool URL, accepted shares, and rejected shares. If a unit only exposes a web UI, leave api_port unset and use web_port for reachability.

CPU Dev Miner

Edit config.toml or the release-provided start-mining.toml:

pool = "stratum+ssl://YOUR_KASPA_POOL_HOST:5555"
wallet = "kaspa:your_wallet_address"
worker = "kaspa-rig-01"
threads = 8
batch_size = 4096
reconnect_secs = 5

Start the TUI:

./target/release/kaspa-miner

From a downloaded release, you can mine directly from the starter config after editing the wallet and pool:

kaspa-miner --config start-mining.toml

Use plain logs:

./target/release/kaspa-miner --no-tui

Override config from the command line:

./target/release/kaspa-miner \
  --pool stratum+ssl://pool.example.com:5555 \
  --wallet kaspa:your_wallet_address \
  --worker rig-01 \
  --threads 8 \
  --batch-size 4096 \
  --no-tui

Supported pool URL schemes:

• stratum+tcp://host:port
• stratum://host:port
• tcp://host:port
• stratum+ssl://host:port
• ssl://host:port

Benchmark And Autotune

Single benchmark:

./target/release/kaspa-miner --benchmark --threads 8 --batch-size 4096 --bench-seconds 15

Autotune:

./target/release/kaspa-miner --tune --tune-max-threads 8 --tune-seconds 5

Custom sweep:

./target/release/kaspa-miner \
  --tune \
  --tune-max-threads 12 \
  --tune-batches 1024,4096,16384,65536 \
  --tune-seconds 8

Tuning rules:

• Start with physical performance cores.
• Increase batch_size to reduce bookkeeping overhead.
• Lower batch_size if you need faster work refresh on high-latency pools.
• Keep clocks stable. Thermal throttling eats hashrate.
• Use nearby pool endpoints to reduce stale shares.
• Use --no-tui for systemd, launchd, Docker, tmux, and log collectors.

CLI Reference

--config <PATH>       CPU miner config path, default: config.toml
--pool <URL>          stratum+tcp://host:port or stratum+ssl://host:port
--wallet <ADDRESS>    kaspa: or kaspatest: address
--worker <NAME>       Worker name appended to wallet for pool login
--threads <N>         CPU worker threads
--batch-size <N>      Nonces per thread before checking pool work
--no-tui              Plain terminal logs
--benchmark           Offline kHeavyHash benchmark
--bench-seconds <N>   Benchmark duration
--tune                Sweep CPU settings and rank local hashrate
--tune-seconds <N>    Per-test duration for --tune
--tune-max-threads N  Maximum thread count for --tune
--tune-batches LIST   Comma-separated batch sizes for --tune
--gpu                 Run managed Kaspa GPU engine mode
--gpu-config <PATH>   GPU engine config path, default: gpu.toml
--gpu-once            Run GPU engine once without watchdog restart
--gpu-info            Print local GPU/runtime discovery
--fleet               ASIC fleet controller mode
--fleet-config <PATH> Fleet config path, default: fleet.toml
--fleet-once          Poll fleet once and exit

Stratum Notes

KASPilot expects the common two-parameter Kaspa notify payload:

mining.notify params: ["jobId", "headerHash"]

The headerHash is parsed as:

32 bytes pre_pow_hash || 8 bytes timestamp_le

The miner submits:

mining.submit params: ["wallet.worker", "jobId", "nonce"]

The nonce is the full 8-byte hex nonce, including any pool-provided extranonce prefix.

Production Checklist

• Confirm your pool supports Kaspa Common Stratum.
• Prefer stratum+ssl:// when the pool supports TLS.
• Keep wallet addresses in config.toml or start-mining.toml, not shell history.
• Keep ASIC management ports on a trusted LAN or VPN.
• Do not expose CGMiner API ports to the public internet.
• Use --gpu to supervise an installed optimized GPU engine; keep its binary updated from a source you trust.
• Use --fleet for ASIC operations and CPU mode for protocol validation.
• Use release builds for benchmarks and production binaries.
• Use RUSTFLAGS="-C target-cpu=native" only when the binary will run on the build machine.
• Watch rejected shares. Repeated low-difficulty, duplicate, or stale share errors usually indicate pool, protocol, or latency issues.

Release Flow

CI runs on Linux, macOS, and Windows. Release archives are produced whenever a v* tag is pushed.

git tag v0.1.1
git push origin v0.1.1

The release workflow builds:

• x86_64-unknown-linux-gnu
• x86_64-pc-windows-msvc
• aarch64-pc-windows-msvc
• macos-universal.pkg

Roadmap

• Vendor-specific adapters for richer IceRiver and Bitmain-style telemetry.
• Native CUDA/OpenCL kHeavyHash backend behind --gpu-native.
• JSON and CSV fleet export.
• Alert thresholds for offline rigs, high temperature, fan faults, and reject spikes.
• Pool failover visibility and worker grouping.
• Optional local web dashboard.
• Apple notarization and Windows Authenticode signing.

References

• Kaspa Common Stratum Protocol: https://file1.iceriver.io/protocols/KAS-Miner-Mining-Protocol-EN.pdf
• Rusty Kaspa PoW primitives: https://docs.rs/kaspa-pow
• Kaspa hash primitives: https://docs.rs/kaspa-hashes
• CGMiner-compatible API pattern: https://docs.luxor88.com/firmware/api/cgminer/summary

License

Cargo.toml declares MIT OR Apache-2.0. Add matching license files before public distribution.