ionews
AI Art

Generative art, from prompt to pipeline

Image and video generation at three depths: no code browser tools, intermediate local setups like ComfyUI, and advanced training of your own models. Costs labelled honestly. The theory behind all of it lives on the Learn page.

Suggested curricula

First images, today no code

Anyone — no code, no GPU · a weekend

From zero to images you're genuinely happy with, using only free browser tools. The skill is seeing and describing, not software.

  1. Generate the same prompt in two tools — Gemini and Grok Imagine are both free — the differences teach you what models actually do
  2. Learn prompt anatomy — subject, medium, style, lighting, composition — Midjourney's docs teach it best, and it transfers everywhere
  3. Iterate, don't re-roll — change ONE thing per generation and watch what moves; ten deliberate iterations beat a hundred lucky dips
  4. Try text and typography in Ideogram — posters and titles are where most models fail; see what a specialist does differently
  5. Animate a favourite still — image-to-video in Grok Imagine or Kling — the fastest "wow" in this field

The ComfyUI path intermediate

Comfortable installing software; GPU with 8GB+ VRAM (or patience) · ~6 weeks at 4 hrs/week

Own your pipeline: unlimited local generation, community models, repeatable workflows. Node graphs look intimidating for about a week, then become the point.

  1. Install ComfyUI (desktop app) — the installer handles Python and models; git only if you like git
  2. Official first-generation tutorial — your first local image, and a map of what the default nodes mean
  3. ComfyUI Wiki basics course — learn what each node does so graphs stop being magic
  4. Pull two community models from Civitai — one photoreal, one stylised; compare against the same prompts from your no-code days
  5. Rebuild three official example workflows — img2img, inpainting, then a Wan video workflow — rebuilding beats downloading

Train your own style advanced

ComfyUI path done; 12GB+ VRAM ideal · ~1 month of evenings

A LoRA that reliably produces YOUR style, character, or subject. Dataset quality decides everything; the training itself is the easy part.

  1. Read the sanj.dev LoRA guide end to end — before touching a trainer — most failed LoRAs die at dataset prep
  2. Build a 20–40 image dataset — consistent subject, varied everything else; caption honestly
  3. First run in FluxGym — the simplest trainer that produces real results; iterate on the dataset, not the knobs
  4. Graduate to ai-toolkit — when you need newer models or finer control
  5. Understand what you just did — the HF course + the diffusion-theory shelf on the Learn page close the loop

Tools & guides

No-code image generation no code 7

Type a prompt, get an image — nothing to install. The assistants you may already use (ChatGPT, Gemini, Grok) now generate images well; the dedicated tools below give more control.

ChatGPT images free tier — GPT Image 2 inside ChatGPT — currently tops blind-vote arenas and the best at legible text in images.
Gemini (Nano Banana) free tier — Google's image models in the Gemini app — Nano Banana Pro renders up to 4K with class-leading text handling.
Grok Imagine free tier — xAI's generator inside Grok: fast image generation, editing, and one-click animation of stills into video.
Midjourney paid — The aesthetic benchmark for years running; V8.1 is fast and offers 4x-resolution HD mode. No free tier.
Ideogram free tier — The typography specialist — when the image must contain real, correctly-spelled words (posters, logos), start here.
Krea free tier — Multi-model creative suite (60+ models plus its own) with a real-time canvas and hosted LoRA training — a gateway toward the intermediate tier. (daily free compute, no card)
Leonardo.Ai free tier — Strong for game assets and stylised work, with the most generous daily free allowance of the major platforms.
No-code video generation no code 7

Text-to-video and image-to-video in the browser. Clips are short (5–15s) and credits run out fast on free tiers — but the quality in 2026 is startling. (OpenAI's Sora was discontinued in April 2026; it's not on this list for a reason.)

Google Veo 3.1 / Flow free tier — The current all-around quality leader: native audio, up to 4K, directed through the Flow filmmaking tool or the Gemini app.
Kling AI free tier — Kuaishou's generator — Kling 3.0 does unified video+audio, native 4K/60fps, multi-shot storyboards up to 15s. Also generates images. (daily credits)
Grok Imagine (video) free tier — Imagine Video 1.5 leads the image-to-video arena: 6–15s clips with native audio, generated very fast.
Runway paid — The professional's pick — Gen-4.5 with camera controls, motion brush, and reference-driven character consistency. (small one-time free credits)
Hailuo AI free tier — MiniMax's generator, notable for physics and photorealism at 1080p.
Luma Dream Machine free tier — The Ray3 model family with the strongest HDR and color handling — but silent output only (no native audio).
Pika free tier — Creator-focused and playful: effect tools like Pikaffects and lip-sync Pikaformance rather than cinematic realism.
Prompting & craft no code 3

Generating a good image is a skill: describing subject, style, lighting, and composition precisely. These teach the craft that transfers across every tool above.

Midjourney documentation free — The best-written prompting reference in the field — parameters, style control, multi-prompts. Worth reading even if you use other tools.
Stable Diffusion Art free — Practical, screenshot-heavy tutorials: prompt anatomy, ControlNet, inpainting, model comparisons.
xAI Imagine docs free — Official capability reference for Grok Imagine — what the model can and can't be steered to do.
Local & node-based — own your pipeline intermediate 7

Some setup required: a decent GPU, an installer or a git clone. In exchange: no credits, no content filters beyond the model's own, total control, and workflows you can save and share.

ComfyUI free — The standard node-based engine for local image and video generation (FLUX, SDXL, Wan). Now ships a desktop installer alongside the classic git install.
ComfyUI — first generation tutorial free — The official zero-to-first-image walkthrough — the right first hour with ComfyUI.
ComfyUI Wiki basics course free — Community course that explains what the nodes actually do, so you can build workflows instead of downloading them.
ComfyUI official example workflows free — Working reference workflows for every major model and task — including the Wan and HunyuanVideo video pipelines.
Invoke free — Artist-oriented alternative to ComfyUI: unified canvas, layers, and inpainting in a polished UI. Actively developed (unlike the dormant Forge).
Civitai free tier — The dominant hub for community checkpoints and LoRAs — browse styles, download models for local use. Mind the licenses per model.
DaVinci Resolve free — The serious free video editor for assembling generated clips; note most AI Neural Engine features need the one-time Studio license. (Studio $295 unlocks AI features)
Train your own — LoRAs & beyond advanced 5

Teaching a model your style, character, or product. A LoRA trains in hours on a consumer GPU; the tooling below is what the community actually uses in 2026.

FluxGym free intermediate — Dead-simple web UI for FLUX LoRA training on 12–20GB VRAM — the gentlest on-ramp to training.
ostris/ai-toolkit free advanced — The current go-to trainer for the newest models (FLUX.2 era); config-file driven, actively maintained.
Kohya_ss GUI free advanced — The long-standing standard trainer for SD 1.5 / SDXL / FLUX — most guides online assume it.
LoRA Training Guide (sanj.dev) free advanced — A current, honest end-to-end guide: dataset prep, captioning, VRAM tricks, and hyperparameters that matter.
Hugging Face Diffusion Models course free advanced — Build diffusion models in code with the diffusers library — for understanding the machinery, not just driving it.

AI art posts