Freedom251 › AI image generators
modl

What modl is
modl is a free, open-source toolkit for generating and editing images and training LoRAs. It runs locally through a command-line interface on Linux and macOS, with a local web interface launched by `modl serve` for generation, training, output management, and LoRA browsing. The CLI creates PNG images from text prompts and applies natural-language instructions to existing images. Additional image tools include upscaling, background removal, segmentation, and ControlNet preprocessing. For LoRA work, modl prepares datasets, captions images, selects parameters, and executes training on your images. Its model page lists FLUX, Z-Image, Qwen Image, and Stable Diffusion support, resolving dependencies automatically. An MCP server connects to Claude Code, Cursor, or other MCP clients locally or over SSH. ComfyUI models can share a store, while local models use SHA256 paths under `~/modl/store/`. Vast.ai GPU rentals can sync results over SSH.
Who it suits
modl suits people who want local image generation, editing, or LoRA training through a CLI or local web interface. It also fits users connecting image workflows with MCP clients or ComfyUI.
What it does well
- Free and open source under AGPL-3.0
- Runs on Linux, macOS, and the web
- Supports text generation and natural-language editing
- Prepares datasets and runs LoRA training
- Connects with MCP clients and ComfyUI
What to know first
- Character LoRA guide requires an NVIDIA GPU with 10+ GB VRAM
- Vast.ai rental billing continues until the pod is destroyed
- GPU rentals require your own Vast.ai API key
Verdict
modl combines local image generation, editing, and LoRA training in one toolkit. Its CLI, local web interface, model support, and workflow connections make it suitable for users who prefer locally managed assets.
Facts from modl's own pages
- Agent integrations
- The built-in MCP server connects to Claude Code, Cursor, or any MCP client locally or over SSH. · modl.run, 12 Oct 2026
- Cloud GPU
- Users can rent GPUs through Vast.ai using their own API key; Vast.ai bills for the rented instance until it is destroyed. · modl.run, 12 Oct 2026
- ComfyUI integration
- The documentation says `modl system link` can adopt recognized ComfyUI models into a shared store so both tools use the same files. · modl.run, 12 Oct 2026
- Creator
- The homepage says modl was built with Rust by Pedro and links to his personal site. · modl.run, 12 Oct 2026
- Data handling
- The maker describes its GPU pod flow as having no hosted middleman or cloud storage, with artifacts syncing to the local library over SSH. · modl.run, 12 Oct 2026
- Free and open source
- The maker describes modl as free and open source, and says local use is free forever. · modl.run, 12 Oct 2026
- Free plan
- Yes · modl.run, 19 Sept 2026
- GPU rental
- Users can rent GPUs through Vast.ai using their own API key, with outputs synced back to their local library over SSH. · modl.run, 12 Oct 2026
- Hardware requirements
- The maker's character LoRA guide calls for an NVIDIA GPU with 10+ GB VRAM and recommends 16–24 GB. · modl.run, 12 Oct 2026
- Image editing
- Yes · modl.run, 19 Sept 2026
- Image generation
- The CLI generates images from text prompts and edits existing images with natural-language instructions. · modl.run, 12 Oct 2026
- Image tools
- It includes image generation, natural-language editing, LoRA training, upscaling, background removal, segmentation, and ControlNet preprocessing. · modl.run, 12 Oct 2026
- Installation
- The maker provides a shell installer and a Cargo install-from-source option. · modl.run, 12 Oct 2026
- License
- The project is free and open source and is licensed AGPL-3.0. · modl.run, 12 Oct 2026
- Local storage
- The docs say models are stored by SHA256 hash in ~/modl/store/. · modl.run, 12 Oct 2026
- LoRA training
- modl trains LoRAs on users’ images and handles dataset preparation, captioning, parameter selection, and training execution. · modl.run, 12 Oct 2026
- Models
- The maker’s model page lists support for models including FLUX, Z-Image, Qwen Image, and Stable Diffusion, with model dependencies resolved automatically. · modl.run, 12 Oct 2026
- Output formats
- PNG · modl.run, 19 Sept 2026
- Purpose
- modl is a local-first toolkit for generating and editing images and training LoRAs from a CLI and web interface. · modl.run, 12 Oct 2026
- Reference images
- Yes · modl.run, 19 Sept 2026
- Rental cost
- The homepage gives an example RTX 3090 rate of $0.209/hr all-in and says Vast.ai billing continues until the pod is destroyed. · modl.run, 12 Oct 2026
- Supported operating systems
- The documentation names systemd on Linux and launchd on macOS for installing the web UI as a system service. · modl.run, 12 Oct 2026
- Text-to-image
- Yes · modl.run, 19 Sept 2026
- Web interface
- The `modl serve` command launches a local web interface for generation, training, output management, and LoRA browsing. · modl.run, 12 Oct 2026
Where modl stands on Freedom251
- AI image generators#48 of 124
