Everything you need to run AI with local control

Install, run, and manage open-source AI tools on your hardware. Local workloads need no cloud account or API keys.

47
AI Apps
41
Model Packs
7
GPU Vendors
100%
Works Offline

Run Any AI App Locally

Install ComfyUI, Stable Diffusion, Open WebUI, and more with one click. Local workloads run on your hardware without API keys or usage fees.

  • One-click install with automatic dependency setup
  • Pre-configured model packs — download once, run anywhere
  • 47 apps across image, audio, video, LLM, and workflow categories

Read our getting started guide →

Works Offline, Syncs When Ready

Your projects, settings, and workflows live on your device first. When you connect a second device, changes merge automatically — no conflicts, no overwrites.

  • Full functionality without internet
  • Automatic conflict-free sync across devices
  • You control what syncs and what stays local

Local Controls, Encrypted Transport

Local data stays on your device until you enable sync or cloud features. Supported network connections protect data in transit; private E2EE is planned.

  • Sync and cloud features are opt-in
  • Encrypted transport for supported network connections
  • Private E2EE is planned, not part of the current release

Local Free, Cloud When You Need It

Run everything locally at no cost. Need more GPU power for large models? Spin up a cloud instance from the same interface.

  • Unlimited local usage — completely free
  • On-demand cloud GPUs for heavy workloads
  • One interface for local runs and cloud deployment controls

No Account Required

Start using the app immediately — no sign-up, no email. Add a recovery phrase when you’re ready. Or connect your existing login.

  • Anonymous by default — just download and run
  • Optional recovery phrase — you hold the keys
  • Sign in with Google or GitHub if you prefer

Every GPU, Auto-Detected

Thumper detects your GPU and configures everything automatically. NVIDIA, AMD, Intel, Apple Silicon — even NPUs on newer laptops.

  • NVIDIA CUDA, AMD ROCm, Intel oneAPI, Apple Metal
  • Automatic VRAM-based model selection
  • NPU acceleration on supported hardware

AI Assistant Built In

A built-in AI assistant with 28 specialized tools for GPU diagnosis, manifest repair, model management, and workflow automation.

  • Natural language interface for complex operations
  • GPU-specific troubleshooting and dependency fixing
  • Automatic manifest generation and validation

Cross-App Image Gallery

Automatically collects generated images from all installed apps with full metadata extraction — prompts, seeds, models, and samplers.

  • Real-time file watching across all app output directories
  • Metadata extraction from ComfyUI, A1111, and Fooocus formats
  • Search and filter by prompt, model, date, or source app

Multi-Device Sync Controls

Pair devices and sync through the server-assisted service. Network traffic uses encrypted transport; private E2EE remains planned.

  • Pairing and authorization controls
  • Automatic CRDT sync across paired devices
  • Manage connected devices from one place

App Reviews & Ratings

Community reviews and star ratings help you find the best apps. Rate apps you’ve used and help others make informed choices.

  • 5-star rating system with written reviews
  • Review aggregation and sorting
  • Verified ‘installed’ badges on reviews

See it in action

Download for free — no account required. Start running AI apps in minutes.