Examples Gallery

Copy-paste starter manifests for common apps, models, and packs. Each example shows the essential fields — expand them with the Manifest Field Reference.

App Manifests (.thumper.yaml)

Minimal App

Minimal
pythonstarter
yaml
thumper: 1
id: my-app
name: My App
version: 1.0.0
description: A simple demo
author: you
type: app
runtime:
type: python
entry: app.py
port: 8080

Open WebUI

Standard
chatollamadocker
yaml
thumper: 1
id: open-webui
name: Open WebUI
version: 0.5.0
type: app
area: chat
runtime:
type: docker
entry: docker-compose.yml
port: 3000
preferred_engine: ollama

ComfyUI

Advanced
imagediffusiongpupatches
yaml
thumper: 1
id: comfyui
name: ComfyUI
type: app
area: image
runtime:
type: python
entry: main.py
port: 8188
gpu_required: true
pack_refs: [sdxl-starter]

ACE-Step

Advanced
audiomusicgpubf16
yaml
thumper: 1
id: ace-step
name: ACE-Step Music
type: app
area: audio
runtime:
type: python
entry: handler.py
port: 7860
gpu_required: true
env:
MIOPEN_FIND_MODE: "2"

Fooocus

Standard
imagediffusionsimple
yaml
thumper: 1
id: fooocus
name: Fooocus
type: app
area: image
runtime:
type: python
entry: entry_with_update.py
port: 7865
gpu_required: true

Model Manifests (.model.yaml)

SDXL Base 1.0

Standard
diffusion-unetsafetensors6.5 GB
yaml
schema: model/v1
id: sdxl-base-1.0
name: SDXL Base 1.0
kind: diffusion-unet
format: safetensors
source:
hf_repo: stabilityai/stable-diffusion-xl-base-1.0
hf_file: sd_xl_base_1.0.safetensors

SDXL VAE FP16

Minimal
vaesafetensors335 MB
yaml
schema: model/v1
id: sdxl-vae-fp16
name: SDXL VAE FP16 Fix
kind: vae
format: safetensors
source:
hf_repo: madebyollin/sdxl-vae-fp16-fix
hf_file: diffusion_pytorch_model.safetensors

OpenVoice Converter

Standard
voice-conversionbin560 MB
yaml
schema: model/v1
id: openvoice-converter
name: OpenVoice v1 Converter
kind: voice-conversion
format: bin
source:
hf_repo: myshell-ai/OpenVoice
hf_file: checkpoints/converter/checkpoint.pth

Pack Manifests (.pack.yaml)

SDXL Starter Pack

Standard
image-generation2 models~7 GB
yaml
schema: pack/v1
id: sdxl-starter
display_name: SDXL Starter Pack
task_category: image-generation
models:
- ref: sdxl-base-1.0
- ref: sdxl-vae-fp16

OpenVoice v1 Base

Standard
voice2 models~1 GB
yaml
schema: pack/v1
id: openvoice-v1-base
display_name: OpenVoice v1 Base
task_category: voice
models:
- ref: openvoice-converter
- ref: openvoice-en-speaker
Use these examples as starting points. Copy one that’s close to what you need, then customize it using the Manifest Field Reference.

Key Takeaways

  • Every app manifest needs at minimum: thumper, id, name, version, type, and runtime
  • Model manifests describe single weight files with schema, id, name, kind, format, and source
  • Pack manifests bundle models into installable units via models[].ref
  • Use gpu_required: true for GPU-dependent apps
  • Copy an example and customize — don’t start from scratch