Overview

Not sure where to start? This page organizes the docs by what you want to accomplish. Pick a task below and we’ll point you to the right guides, tutorials, and reference pages.

Tasks are grouped into three tracks: Create (use AI apps to generate content), Develop (build and package AI apps), and Deploy (sync, self-host, and migrate).

Create

Use Thumper-Run to generate images, chat with AI, create music, or produce video — all running locally on your hardware.

Generate Images

Set up ComfyUI with SDXL and create your first AI-generated image in minutes.

Tutorial: First Image →

Chat with AI

Run a local LLM with Open WebUI and Ollama for private, offline conversations.

Tutorial: First Chatbot →

Generate Music

Create AI-generated music tracks with ACE-Step, optimized for your GPU.

Tutorial: First Music →

Generate Video

Produce short AI video clips using LTX-Video through ComfyUI workflows.

Develop

Package AI tools for the catalog, create model packs, build custom nodes, or configure offline model access.

Package an App

Write a .thumper.yaml manifest to package any AI tool for the catalog.

Publishing Guide →

Create Model Packs

Bundle models with optimized configs for different GPUs using .model.yaml and .pack.yaml.

Manifest Reference →

Build Custom Nodes

Extend ComfyUI with custom Python nodes that integrate with Thumper workflows.

Run Models Offline

Download and configure models for fully offline AI inference with no internet required.

User Guide: Models →

Deploy

Sync your setup across devices, run your own server, or migrate from another platform.

Sync Across Devices

Pair devices for supported server-assisted sync workflows. Private E2EE remains planned.

User Guide: Pairing →

Self-Host the Platform

Deploy your own Thumper-Run server with Docker, PostgreSQL, and OIDC auth.

Migrate from Pinokio

Move your existing Pinokio apps and models to Thumper-Run with minimal friction.

Key Takeaways

  • Create — generate images, chat, music, or video with local AI apps
  • Develop — package apps, create model packs, build custom nodes
  • Deploy — sync across devices, self-host, or migrate from Pinokio