AI tutorials
Learn how generative AI works and how to get good results from it. Each guide answers the question first, then shows the method with real prompts and examples, the common mistakes and the official sources.
AI basics
- 7 min readWhat is generative AI?How AI models learn from data to make text, images, video and audio, and where they fail
- 7 min readHow LLMs workFollow one prompt through tokens, attention and next-token prediction to an answer
- 6 min readTokens and context windowReal token counts in six languages, what fills the context window, and how to save space
- 6 min readAI hallucinationWhy AI invents facts and citations, prompts that reduce it, and how to check answers
- 8 min readHow diffusion models workHow noise becomes an image, and what steps, seed and CFG scale change
- 7 min readWhat are AI agents?How AI agents call tools in a loop, a real trace, MCP, and the risks to plan for
- 8 min readAI glossary60 generative AI terms in plain English, from tokens and LoRA to CFG, C2PA and deepfakes
Prompt writing
- 6 min readPrompt engineering basicsThe six parts of a good prompt, a before/after rewrite, examples and testing
- 7 min readHow to write image promptsA prompt formula, a weak-to-strong rewrite, keyword tables and negative prompts
- 7 min readHow to write video promptsA 7-part prompt structure, one action per shot, beat timing and a full rewrite
- 7 min readCamera movement termsDolly, truck, orbit, crane, dolly zoom and more: what each looks like and how to prompt it
- 8 min readStoryboard promptsTurn a script into shots and one prompt per shot, with a 15-second example.
AI images
- 7 min readHow text-to-image AI worksFrom prompt to picture: steps, guidance, seed, size and a working draft-to-final flow
- 7 min readHow image-to-image AI worksStrength, reference images, style transfer, inpainting and depth, pose and edge control
- 7 min readConsistent AI charactersCharacter sheets, references, seeds, fixed descriptors, LoRA and a continuity checklist
- 6 min readHow AI upscaling worksWhy bicubic blurs, how GAN and Real-ESRGAN models add detail, and when they fail
- 7 min readHow AI background removal worksMasks vs alpha mattes, why hair and glass fail, and why cut-outs must be PNG
- 6 min readWhat is a depth map?How AI guesses depth from one photo, near vs far shading, and uses for 3D and ControlNet
AI video
- 6 min readHow text-to-video worksHow a prompt becomes a clip, the settings that matter, and a 15-second shot plan
- 6 min readHow image-to-video worksPick a start image, write motion-only prompts, and when 2.5D parallax is enough
- 6 min readFirst and last frame videoKeyframe interpolation, how to design both frames, and transitions that work or fail
- 7 min readMulti-shot AI videoKeep faces, clothes, places and light consistent across many AI video shots.
- 8 min readAI short drama workflowFrom hook and script to a labeled 9:16 episode, with a 20-second shot list.
AI audio
AI for developers
Copyright and safety
- 7 min readAI copyright and commercial useWho owns AI output in the US, China and EU, and what to check before selling it.
- 7 min readHow to label AI contentExplicit and implicit labels under China's rules, EU AI Act and platform policies.
- 7 min readC2PA and AI watermarksContent Credentials, SynthID and IPTC labels, why metadata vanishes, and detection limits
The basics explain what large language models and diffusion models actually do, why they hallucinate and what tokens and context windows are. The prompt guides give reusable formulas for text, image and video prompts, with weak and strong versions side by side. The image, video and audio guides walk through each workflow from the first prompt to a finished file, and say where the tools still fail.
The guides are model-agnostic, so the methods work in ChatGPT, Claude, Gemini, Midjourney, Stable Diffusion, Kling, Veo and similar tools. Legal pages cite the official texts and are general information, not legal advice. Several guides link to cv.cm's free AI tools, which run in your browser without uploading your files.
FAQ
Where should a beginner start with AI?
Read what generative AI is and how LLMs work, then the prompt writing basics. Those three pages explain most of what you need to get useful answers from any chatbot or image generator.
Do these tutorials depend on one AI tool?
No. They teach methods that work across tools. Where a step depends on a product, the guide says so and describes the general option to look for.
Can I use AI-generated images and videos commercially?
Often yes, but it depends on the tool's license and terms, the law where you publish, and whether the output copies someone else's work or likeness. The copyright and labeling guides list what to check.
Are the cv.cm AI tools free?
Yes. The AI image generator, photo to video, background remover, upscaler and speech to text tools are free and run on your own device, so your files are not uploaded.