YAML Metadata Warning:empty or missing yaml metadata in repo card

Check out the documentation for more information.

HWcontrol2.0

HWcontrol2.0 is a cross-platform PC hardware monitoring and control application for Windows 10/11, Linux, and macOS 14+ Apple Silicon. It is designed to report CPU/GPU temperatures, memory and storage telemetry, fan speeds and other sensors when the operating system and hardware expose them; supported devices can also use safety-gated fan and performance controls.

It is a desktop PC utilityโ€”not an industrial automation/SCADA product, a serial hardware-flow-control protocol, or a DAW/mixer controller. Its architecture is Flutter/Dart desktop GUI โ†’ authenticated localhost Go bridge โ†’ native C++/CMake engine โ†’ platform-specific hardware APIs. Hardware support varies by device, driver, operating system, and available permissions; this project is in Beta.

HWControl Website GitHub Releases GitLab Mirror Codeberg Mirror GitHub Repository

HWcontrol2.0 is an open-source, cross-platform hardware monitoring, control, health and diagnostics platform built around a Flutter/Dart dashboard, an authenticated localhost Go bridge, and a native C++/CMake engine.

Current channel: Beta. Hardware support and control capabilities depend on the operating system, drivers, and detected hardware. Unsupported control paths remain monitor-only.

Beta download

The current Beta distribution uses platform-specific GitHub Release channels:

Release assets are generated by GitHub Actions and accompanied by SHA-256 manifests when the platform release workflow completes successfully. Do not treat a release as downloadable until its expected package assets are present on the release page.

What it does

  • CPU, memory, SSD/HDD/NVMe, GPU and fan telemetry where exposed by the OS/driver
  • Per-volume storage usage, free capacity and read/write throughput; device health is reported only when a trusted platform-specific health source is available
  • Temperature history and configurable temperature alerts
  • Hardware-control commands for supported platforms/devices
  • Performance / Balanced / Quiet profile support where available
  • Game Mode
  • Safety-normalized Smart Fan Curve editor with a nonzero cooling floor, thermal emergency override and capability-gated automatic control
  • Per-channel fan inventory, independent manual targets, local fan profiles, and CPU/GPU/hottest-sensor selection where the platform exposes validated PWM channels
  • Local temperature automation rules
  • Bounded CPU microbenchmark
  • Local extension metadata/capability contract
  • Local user profiles and fan-curve configuration
  • Secure loopback communication between the dashboard and bridge
  • Bridge diagnostics, hardware capability matrix, anomaly detection and fail-safe health states
  • Persistent telemetry history and local profile storage
  • Optional local AI support; the core application does not depend on the model being installed
  • Verified GitHub Release update download with SHA-256 validation when release asset digests are available
  • Cross-platform build, package, checksum and provenance validation in GitHub Actions

Supported release targets

Platform Release target Package formats Architecture Status
Windows Windows 10 / 11 EXE + MSI + MSIX + portable ZIP x64 Beta
macOS macOS 14+ PKG + DMG containing .app Apple Silicon / arm64 Beta
Linux Debian/Ubuntu-family DEB + TAR.GZ + TAR.ZST x64 Beta

Windows ARM64 and Intel macOS packages are not currently shipped.

Installation

Windows

Use the guided EXE installer for normal users. The MSI remains available for managed deployment, MSIX for modern Windows package deployment, and the portable ZIP for users who do not want an installed Windows service.

macOS

Use the Apple Silicon PKG or DMG from the macOS Beta release.

Linux

Use the DEB on Debian/Ubuntu-family systems, or the portable TAR package when appropriate.

Debian/Ubuntu APT repository (amd64)

HWcontrol2.0 also publishes a signed APT repository through GitHub Pages. The repository setup script pins the archive signing-key fingerprint and refuses to install a key that does not match it.

Download and inspect the setup script, then register the repository:

curl --fail --location --proto '=https' --tlsv1.2 \
  https://raw.githubusercontent.com/lordkeremello45/HWcontrol2.0/main/scripts/setup-apt-repository.sh \
  --output setup-hwcontrol-apt.sh
less setup-hwcontrol-apt.sh
sudo bash setup-hwcontrol-apt.sh

After the script finishes successfully, install or update HWcontrol with:

sudo apt install hwcontrol

The setup script installs the repository key into /usr/share/keyrings/, writes /etc/apt/sources.list.d/hwcontrol.sources, and runs apt-get update. It supports Debian/Ubuntu-family amd64 systems only. The APT repository currently selects stable GitHub Releases; Beta-only builds are not automatically published to this channel. Repository registration does not guarantee that a package is available: if no stable Linux DEB release is published, installation will fail until one is available.

For detailed platform instructions, see the Wiki. For a careful first install, checksum guidance and issue-reporting checklist, use the Beta Quickstart.

Why this project โ€” and an honest status report

We want HWcontrol2.0's strengths and limitations to be easy to evaluate, not hidden behind marketing claims:

These documents distinguish project goals from verified capabilities. In particular, CI success is not a substitute for a documented virtual-machine and physical-hardware test matrix.

User help and product readiness

The checklist is an acceptance standard, not a claim that every gate has passed. Use real CI, packaging and hardware-test evidence to determine readiness for each platform.

Language and localization

HWcontrol2.0 provides a System Default language option. The Language Service reads the operating system's current locale and resolves it against the localization resources bundled with the application. Manually selected languages are stored in the user's local settings and take precedence over the system locale.

When System Default is selected, the application re-evaluates the operating-system locale on startup. If the exact locale is not bundled, HWcontrol2.0 uses the matching language resource when available; otherwise its existing English fallback remains in effect.

Security and release integrity

The local bridge listens on loopback and authenticates commands with HMAC-SHA-256. Release workflows publish SHA-256 manifests and GitHub artifact attestations where configured.

Checksums and attestations provide integrity/provenance evidence; they are not substitutes for OS-trusted publisher signing. Windows Authenticode and Apple Developer ID/notarization are only claimed when the corresponding trusted credentials and verification steps are actually present.

See SECURITY.md and the security documentation.

Ada/SPARK security core

HWcontrol2.0 includes an Ada package using SPARK (security/spark-core/src/hwcontrol_security.ads) for contract-based validation of safety-sensitive fan commands. The core defines sensor-temperature bounds and checks command authorization, telemetry validity, integrity state, and critical-temperature conditions before returning an allow/deny decision.

  • SPARK source: security/spark-core/
  • CI workflow: SPARK security core, configured to build, test, and run formal verification.
  • Scope: this is a security-policy component; it does not by itself prove that every GUI/Go Bridge/native-engine command is routed through it at runtime.
  • Verification status: consult the linked GitHub Actions runs for the current result. The presence of SPARK contracts or a workflow is not, on its own, evidence that formal proof passed or that end-to-end enforcement has been validated.

Architecture

Flutter / Dart GUI โ†’ Authenticated localhost Go Bridge โ†’ Native C++ / CMake Engine โ†’ Platform-specific hardware APIs

The native engine provides the low-level hardware integration layer; the overall project covers monitoring, control, health, diagnostics, safety and release integrity.

NVIDIA GPU development

The Go bridge can use nvidia-smi for NVIDIA GPU telemetry when the installed driver exposes the requested fields. This is separate from GPU-accelerated inference.

The native llama.cpp build supports an opt-in CUDA configuration for local AI development. CUDA is disabled by default so ordinary cross-platform builds do not require the CUDA toolkit. On a supported NVIDIA development machine with a compatible CUDA toolkit installed:

cmake -S . -B build-cuda -DHWCONTROL_ENABLE_CUDA=ON
cmake --build build-cuda --config Release

This option has been added to the build configuration, but CUDA-enabled builds and runtime performance must still be validated on actual supported NVIDIA hardware before claiming release support. Standard release artifacts should not be assumed to use CUDA.

Official references:

Development

The repository validates the native engine, Go bridge, Flutter dashboard, packaging paths and release integrity through GitHub Actions.

For contributors:

Beta feedback

Found a bug, hardware compatibility issue, or have a suggestion? Please use the HWcontrol2.0 Beta Feedback Form.

Beta feedback is intentionally routed through Google Forms โ†’ Google Sheets โ†’ Apps Script โ†’ GitHub Issues. The integration filters contact/private fields, adds duplicate protection and retries, and does not require feedback submitters to expose bridge credentials.

If you report a hardware problem, include the OS, architecture, CPU/GPU model, driver version and relevant telemetry. Never include passwords, API keys, bridge keys, signing credentials or private tokens.

Community and public engineering

HWcontrol2.0 is developed in the open. The public engineering plan focuses on useful technical documentation and reproducible discussion rather than paid promotion.

Licensing

  • Desktop application: GNU GPLv3
  • Website source: GNU AGPLv3-only
  • Local security tooling under security/local-audit/: Apache License 2.0 (Apache-2.0)
  • Third-party dependencies: their respective licenses

See License documentation, LICENSE-APACHE, THIRD_PARTY_NOTICES.md, and Technical References.

Project status

The project is under active Beta development. CI and release automation are treated as part of the product, not as a substitute for real hardware validation. A platform being buildable does not imply that every motherboard, GPU, sensor or fan controller is supported.

Website

The official project website provides platform-specific release selection and resolves GitHub Release assets:

https://lordkeremello45.github.io/HWcontrol2.0/

Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐Ÿ™‹ Ask for provider support