What hardware do I need?

Most modules require Windows 10/11, 16 GB RAM, and a GPU with at least 6 GB VRAM. We support both NVIDIA GPUs (via CUDA) and AMD GPUs (via Vulkan). For the best experience, 32 GB RAM and 8+ GB VRAM are recommended. Eidolon Micro runs on CPU only with just 4 GB RAM.

Does it work with AMD GPUs?

Yes. All GPU-accelerated modules support both NVIDIA (CUDA) and AMD (Vulkan) graphics cards. The installer automatically detects your GPU and configures the appropriate backend.

How do I activate my license?

After purchase, you receive a license key in your order confirmation email and in your account dashboard. Enter the key during installation when prompted. Your license is tied to your hardware fingerprint.

Can I install on multiple machines?

Standard licenses allow installation on 1 machine. Kickstarter backer licenses allow up to 5 machines. If you need to move your license to a different machine, you can remove the old machine from your account dashboard.

Does it work offline?

Yes. Once installed and activated, all modules run entirely on your hardware without an internet connection. The only feature that requires internet is the web search function in Chat and GPT-OSS. License activation requires a one-time internet connection.

Is my data private?

Absolutely. All AI processing happens on your machine. No data is sent to any server. No telemetry, no analytics, no cloud processing. Your conversations, images, documents, and generated content stay on your hardware.

Are there recurring fees or subscriptions?

No. All purchases are one-time. You own the software and can use it indefinitely. There are no monthly fees, no API usage charges, and no per-generation costs.

How do I download my modules after purchase?

Log in to your account at eidolonhub.com. Your purchased modules appear in your dashboard with download links. Each module includes an installer that handles setup automatically.

Do I get updates?

Yes. Updates for your purchased modules are available through your account dashboard. Major version upgrades may be offered at a discounted price for existing customers.

What is a Kickstarter backer discount?

If you backed the Eidolon Home AI Kickstarter campaign, you receive an automatic 10% discount on all new module purchases, and your license allows installation on up to 5 machines instead of 1.

Can I buy individual modules if I already have a pack?

Yes. Individual modules can be added to any existing installation. Simply purchase the module, download it, and run the installer — it integrates with your existing Eidolon setup.

What is your refund policy?

We offer a 14-day refund for undownloaded products. If you’ve downloaded the software and experience issues, contact us — we’ll work to resolve the problem or offer a refund on a case-by-case basis. See our full refund policy.

GPU not detected or “No GPU found” error

NVIDIA users: Make sure you have the latest NVIDIA drivers installed. The installer requires CUDA-capable drivers (version 525+). Download from nvidia.com/drivers.
AMD users: Ensure you have the latest AMD Adrenalin drivers with Vulkan support. The installer uses the Vulkan backend for AMD GPUs. If detection fails, the system will try a 3-level cascade: registry → dxdiag → fallback.

Voice features not working (Piper TTS)

If text-to-speech is not producing audio, check that the Piper TTS binary is correctly located at tools\piper\piper\piper.exe within your Eidolon installation folder. If the file is missing, re-run the installer.

HWiNFO64 monitoring not showing data

For hardware monitoring integration, HWiNFO64 must have “Shared Memory Support” enabled. Open HWiNFO64 → click on the Gadget tab → check “Enable Shared Memory”. Then restart the Eidolon module.

Installation hangs or fails on startup

If the installer appears to hang, check that you have enough disk space (at least 20 GB free for standard modules). Some modules download large AI models during first run — this can take several minutes depending on your internet speed. Let it complete without interrupting.

Out of VRAM error during generation

If you see VRAM-related errors during image or video generation, close other GPU-intensive applications (games, other AI tools, video editing). For video generation modules (Wan, Pyramid Flow), the system uses CPU offloading by default to reduce VRAM usage. If errors persist, try reducing the resolution or batch size in the module’s settings.

License activation fails

License activation requires an internet connection. If activation fails: verify your license key is correct (check your order email or account dashboard), ensure your machine has internet access, and check that you haven’t exceeded your machine limit (1 for standard, 5 for backers). You can remove old machines from your account dashboard.

Module not appearing after installation

Make sure you ran the installer as administrator. Check that the module was installed in the correct Eidolon directory. If the issue persists, try re-running the installer — it will detect the existing installation and repair any missing files.

“cublas64_13.dll not found” or similar DLL missing error

This error means your NVIDIA CUDA runtime libraries are missing or outdated. To fix it:

Option 1 (recommended): Update your NVIDIA GPU drivers to the latest version from nvidia.com/drivers. The latest Game Ready or Studio drivers include the required CUDA libraries.

Option 2: Install the CUDA Toolkit separately from developer.nvidia.com/cuda-downloads. Version 13.x or later is required.

After installing, restart your PC before launching the Eidolon module. If the error persists, check that the CUDA bin folder (typically C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v13.x\bin) is in your system PATH.

GPU not detected — module defaults to CPU mode

If your module starts but reports “No GPU detected” or runs significantly slower than expected, it may be falling back to CPU mode. Check the following:

NVIDIA users: Ensure you have the latest drivers (version 525+). Open a command prompt and run nvidia-smi — if this command is not found, your drivers are not properly installed.

AMD users: Ensure you have the latest AMD Adrenalin drivers with Vulkan support. The system uses a 3-level detection cascade (registry → dxdiag → fallback). If all three fail, it means Vulkan is not available on your system. Try reinstalling drivers with a clean install (use AMD Cleanup Utility first).

General checks: Make sure no other GPU-intensive application is locking the GPU. Close games, video editors, and other AI tools before starting Eidolon. On laptops with dual graphics (integrated + dedicated), make sure Eidolon is set to use the dedicated GPU in Windows Settings → Display → Graphics.

Module running on CPU instead of GPU (very slow generation)

If image, video, or music generation is taking much longer than expected (e.g., 10+ minutes for a single image), the module is likely running on CPU instead of GPU. This can happen even when a GPU is present. Here’s how to diagnose and fix it:

1. Check your GPU VRAM: Open Task Manager → Performance → GPU. Check “Dedicated GPU memory”. If your GPU has less than 4 GB VRAM, some modules will automatically fall back to CPU mode. The minimum for GPU-accelerated generation is 6 GB VRAM.

2. Check the module log: When the module starts, the console or log window shows which backend it is using (CUDA, Vulkan, or CPU). Look for lines mentioning “backend”, “device”, or “GPU”. If it says “cpu” or “fallback”, the GPU was not detected or has insufficient VRAM.

3. Free up VRAM: Close other applications that use GPU memory (browsers with hardware acceleration, video players, other AI tools). Even a web browser can use 500 MB–1 GB of VRAM.

4. Reinstall with GPU support: If you initially installed the CPU-only version, you’ll need to reinstall the module and select the GPU (CUDA or Vulkan) backend during setup.

Note: Some modules like Eidolon Micro and GPT-OSS 120B LITE are designed to run on CPU. For these, CPU mode is normal and expected.