← Back to Skills Marketplace

LFIT

Name: LFIT
Author: jchandler187

by Lowwattlabs · GitHub ↗ · v1.0.3 · MIT-0

cross-platform ✓ Security Clean

Downloads

Stars

Active Installs

Versions

Install in OpenClaw

/install lfit

Description

⚡ LFIT — Local HD image generation via stable-diffusion.cpp on Vulkan/iGPU. Free, private, on your hardware.

README (SKILL.md)

⚡ LFIT — Local Free Image Tool

Local HD image generation. Free. Private. On your hardware.

LFIT runs stable-diffusion.cpp on your Vulkan-capable GPU. No cloud API keys, no per-image fees, no data leaving your machine. Three locked presets — pick by name, get consistent results.

Quick Start

# Install
npm install -g @lowwattlabs/lfit

# Set up the server (first time only)
lfit-server setup

# Start the server
lfit-server start --daemon

# Generate
lfit --preset standard --prompt "a knight at a castle gate"
lfit --preset background --prompt "misty pine valley at dawn"
lfit --preset hero --prompt "ancient dragon, key art" --yes

Presets

Preset	Use for	Size	Time
`standard`	Characters, items, single subjects	1024×1024	~2m30s
`background`	Scenes, environments, wallpapers	1344×768	~2m30s
`hero`	Final/max-quality key art (no LoRA)	1024×1024	~13 MINUTES

LoRA Self-Check

The #1 failure mode: the server ignores \x3Clora:> prompt tags. If LoRA fails silently, you get a blurry base-SDXL image. LFIT detects this two ways:

Small output warning — image data under 1000 bytes
Fast completion warning — standard/background finishing under 30 seconds means LoRA didn't apply

If you see these, check: curl -s http://127.0.0.1:7860/sdapi/v1/loras

Requirements

Python 3.8+
Vulkan-capable GPU (4GB VRAM min, 8GB for hero)
sd-server binary with Vulkan support
SDXL base 1.0 model from HuggingFace
SDXL Lightning 8-step LoRA from HuggingFace

Pricing

Free. If it saves you time, buy us a coffee.

License

MIT-0 — Low Watt Labs ⚡

Usage Guidance

Install only if you are comfortable trusting the external @lowwattlabs/lfit npm package and any server/model downloads it performs. Keep the server bound to 127.0.0.1, avoid lfit-quick if you do not want prompts sent to a public cloud service, and leave Telegram credentials unset or use --no-telegram unless you want generated images sent to Telegram.

Capability Assessment

✓ Purpose & Capability

The stated purpose is local SDXL image generation through stable-diffusion.cpp, and the manifest/config align with a localhost image-generation provider, model directories, LoRA directories, output paths, presets, and timeouts.

ℹ Instruction Scope

The README discloses Pollinations cloud quick drafts and optional Telegram image push, but the privacy wording is overbroad because it also says no data leaves the machine; this is a clarity issue rather than hidden behavior.

ℹ Install Mechanism

Installation is a global third-party npm package and setup may download or use an sd-server binary and model assets. That is coherent for this tool, but the submitted skill artifact itself contains only documentation and metadata, not the referenced executable scripts.

ℹ Credentials

Localhost server access, GPU/model paths, and generated-output storage are proportionate to local image generation. External services are optional but would send prompts or generated images off the machine when used.

ℹ Persistence & Privilege

The skill documents a user-started daemon mode with a stop command, and the manifest has onStartup false; it does not request automatic startup, broad indexing, credential harvesting, or elevated persistence.

How to Use

Make sure OpenClaw is installed (local or Docker)
Run the install command in chat: /install lfit
After installation, invoke the skill by name or use /lfit
Provide required inputs per the skill's parameter spec and get structured output

Version History

v1.0.3

v1.0.3 — Added Low Watt Labs cross-links (Frisk, HOARD)

v1.0.2

v1.0.2 — Tighter stop command: uses lsof to find PID by port (won't kill other sd-server instances), pkill fallback with port-specific regex instead of broad match.

v1.0.1

v1.0.1 — Public README rewrite (removed all Jason-specific references, Node-1, UID 1000, systemd units). Removed duplicate config keys in plugin manifest (serverUrl/loraDir/outDir no longer duplicated at top-level and under image overlay).

v1.0.0

v1.0.0 — LFIT: Local Free Image Tool. Local HD image generation via stable-diffusion.cpp on Vulkan/iGPU. Three presets, LoRA self-check, server launcher, troubleshooting guide.

Metadata

Slug lfit

Version 1.0.3

License MIT-0

All-time Installs 0

Active Installs 0

Total Versions 4

Frequently Asked Questions

What is LFIT?

⚡ LFIT — Local HD image generation via stable-diffusion.cpp on Vulkan/iGPU. Free, private, on your hardware. It is an AI Agent Skill for Claude Code / OpenClaw, with 49 downloads so far.

How do I install LFIT?

Run "/install lfit" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is LFIT free?

Yes, LFIT is completely free, licensed under MIT-0. You can download, install and use it at no cost.

Which platforms does LFIT support?

LFIT is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).

Who created LFIT?

It is built and maintained by Lowwattlabs (@jchandler187); the current version is v1.0.3.

More Skills