This free AI prompt library holds 30 copy-and-paste prompts of this site’s own: seven for images, seven for video, six for writing, six for work and study, and four for coding. Copy one, change the words it names, and paste it into your AI tool.
Each has a USE WHEN label and a “change this” line naming the words to swap for your own.
What is in this prompt library?
The plain point: five groups, each for a different kind of tool.
| Group | What it covers | Check first |
|---|---|---|
| Images (7) | Covers: portrait, product, landscape, food, illustration, a photo edit, a poster. | Check: does your tool read a negative prompt? |
| Video (7) | Covers: one per shot type, a photo brought to life, a line of dialogue. | Check: camera words. |
| Writing (6) | Covers: email, summary, proofreading, post, explanation, outline. | Check: the text is fine to share. |
| Work and study (6) | Covers: meeting actions, comparison, interview, quiz, weekly plan, formula. | Check: the answer against your notes. |
| Coding (4) | Covers: a one-page site, explaining code, fixing an error, accessibility. | Check: safe to paste. |
More: starter prompts for everyday tasks and the prompt vault.
How do you use these prompts?
The plain point: copy, change the named words, paste, then improve one thing at a time. Google gives the same advice for its image model: do not expect a perfect first result, make small changes in conversation.1
The order inside each prompt follows the vendors, and they differ: Black Forest Labs puts the subject first for FLUX.2,2 while Google’s Veo formula starts with the camera.3 All formulas: what the vendors say.
Image prompts: portraits, products, places and edits
The plain point: subject early, then light, lens or style, and what the image is for, which Google asks you to state.4 On ChatGPT Images, things to leave out go inside the prompt;5 on Midjourney they go at the end, after --no.6
Change this: the person, the room, the light.
Change this: the product, the background.
Change this: the place, the time of day.
Change this: the dish, the table.
Change this: the animal. Describe styles in words, never a living artist’s name.
Change this: the one change. More on image edit prompts.
Change this: the event, the words. Check the spelling (text inside images).
Video prompts: one for each kind of shot
The plain point: one shot size, one camera move, one action per clip. Runway advises starting with only the essential motion,7 and Kling puts the subject first, then how it moves, then the scene.8
Change this: the place, the pan direction.
Change this: the person, the feeling.
Change this: who moves, the setting.
Change this: the object, the surface.
Change this: the person, the task.
Change this: the movement. Steps: photo to video.
Change this: the speaker, the line, the sounds. The missing quotation marks are on purpose: Google’s newest Veo guide puts speech after a colon so it is not written on screen.9
Writing prompts: emails, summaries and posts
The plain point: give the goal, the facts and the length, and say what you want rather than what you do not, as Anthropic advises.10
Change this: the person, goal and facts.
Change this: the reader.
Change this: nothing, unless you want clarity help.
Change this: the platform, news, audience.
Change this: the topic.
Change this: the title, reader, length.
Work and study prompts: meetings, plans and quizzes
The plain point: say what shape you want back, and ask for “not stated” instead of guesses. Anthropic’s test: would a colleague with little context follow your prompt? If not, nor will the AI.11
Change this: the columns.
Change this: the columns.
Change this: the job. More on interview prep with AI.
Change this: the question mix.
Change this: your lists and best hours.
Change this: the program and columns. Test on a copy (formulas with AI).
Coding and website prompts
The plain point: say what to build, what it must do and what to leave alone; Anthropic asks for specific output formats and constraints.11
Change this: the purpose. Free hosts: free website tools.
Change this: the language.
Change this: the error, code, version.
Change this: the checks. It is a first pass only.
A library prompt is a starting shape. The words you swap in do most of the work, and your tool’s own formula decides where they go.
Questions people ask
Are these prompts free to use?
Yes. They are this site’s own examples to copy and reuse, as starting points.
Do these prompts work in every AI tool?
The writing, work and coding ones work in any chatbot. Image tools differ: FLUX.2 reads no negative prompt, so describe what you want present instead.12 Which models take a negative prompt lists each.
Why does my result not match the prompt?
Partly because some tools rewrite it: OpenAI’s image API revises your prompt before generating.13 Change one thing and try again; why your generation came out wrong covers the usual causes.
How do I write my own prompt?
Say what you want, who it is for and what shape the answer takes. What makes a good prompt gives six parts; prompt frameworks compared covers memory aids such as CO-STAR.
Where to go next
- “Build one for me.” The prompt generator.
- “What does my tool want?” Its cheat sheet.
- “My images look fake.” Realism words.
- “Teach me the method.” How to write AI prompts, prompting techniques.
- “What do these words mean?” Prompt, text-to-image, text-to-video, image-to-image.
QUICK CHECK · 3 QUESTIONS
Pick an answer to see if you got it. Nothing is recorded.
1. What does the “change this” line under each prompt tell you?
- Which words to swap for your own details
- Which AI tool the prompt was tested in
- How many times to run the prompt
Show the answer
Answer: Which words to swap for your own details. The page says each prompt has a change-this line naming the words to swap for your own.
2. Why do the image and video prompts put their parts in different orders?
- To make the prompts harder to copy
- Because each vendor publishes its own formula, such as subject first for FLUX.2 or camera first for Veo
- Because longer prompts always work better
Show the answer
Answer: Because each vendor publishes its own formula, such as subject first for FLUX.2 or camera first for Veo. The page explains that the order follows the vendors, and their formulas differ.
3. What should you do before pasting code or text into an AI tool, according to the page?
- Translate it into English
- Add a role to the start of it
- Remove names, account numbers, passwords and keys
Show the answer
Answer: Remove names, account numbers, passwords and keys. The page says nothing private goes in, and to check code with the safe-to-paste tool.
- 1Google AI for Developers, Nano Banana image generation, read 16 Sep 2026 and re-read in a browser 13 Oct 2026 (as cited on this site’s Nano Banana cheat sheet): “Don’t expect a perfect image on the first try. Use the conversational nature of the model to make small changes.”↑
- 2Black Forest Labs, FLUX.2 prompting guide, read at source 16 Sep 2026 (as cited on this site’s FLUX cheat sheet): “Subject + Action + Style + Context”.↑
- 3Google Cloud, Ultimate prompting guide for Veo 3.1, read at source 16 Sep 2026 (as cited on this site’s Veo cheat sheet): “[Cinematography] + [Subject] + [Action] + [Context] + [Style & Ambiance]”.↑
- 4Google AI for Developers, Nano Banana image generation, read 16 Sep 2026 and re-read in a browser 13 Oct 2026 (as cited on this site’s Nano Banana cheat sheet): “Explain the purpose of the image. The model’s understanding of context will influence the final output.”↑
- 5OpenAI, Image prompting guide, read at source 16 Sep 2026 and again 13 Oct 2026 (as cited on this site’s ChatGPT Images cheat sheet): “State exclusions such as unwanted text, logos, or watermarks.”↑
- 6Midjourney, No parameter, read in a browser 16 Sep 2026 (as cited on this site’s Midjourney cheat sheet): “Add --no to the end of your prompt in the Imagine bar, followed by the thing or list of things you don’t want in your image.”↑
- 7Runway, Gen-4 Video Prompting Guide, read in a browser 17 Sep 2026 (as cited on this site’s official formulas page): “Begin with a foundational prompt that captures only the most essential motion to the scene.”↑
- 8Kling AI, Text-to-video prompt guide, read at source 16 Sep 2026 (as cited on this site’s Kling cheat sheet): “Subject (description) + Subject Movement + Scene (description) + (Camera Language + Lighting + Atmosphere)”.↑
- 9Google Cloud, Best practices for generating videos, last updated 9 Oct 2026, read at source 13 Oct 2026 (as cited on this site’s Veo cheat sheet): “To prevent the model from rendering text in the video, use a colon ( : ) after the speaker’s action to denote speech and avoid using quotation marks”.↑
- 10Anthropic, Prompting best practices, Claude developer docs, undated, read at source 19 Oct 2026“Tell Claude what to do instead of what not to do”. ↑
- 11Anthropic, Prompting best practices, read at source 19 Oct 2026“Golden rule: Show your prompt to a colleague with minimal context on the task and ask them to follow it. If they’d be confused, Claude will be too.” ↑a“Be specific about the desired output format and constraints.” ↑b
- 12Black Forest Labs, FLUX.2 prompting guide, read at source 16 Sep 2026 (as cited on this site’s FLUX cheat sheet): “FLUX.2 does not support negative prompts. Focus on describing what you want, not what you don’t want.”↑
- 13OpenAI, Image generation guide, read at source 16 Sep 2026 (as cited on this site’s official formulas page): the main model “will automatically revise your prompt for improved performance”.↑
Reused from this site’s cheat sheets and official formulas page, with the read dates recorded there: Black Forest Labs’ FLUX.2 prompting guide (16 Sep 2026); Google’s Ultimate prompting guide for Veo 3.1 (16 Sep 2026) and Best practices for generating videos (13 Oct 2026); Google’s Nano Banana image generation docs (13 Oct 2026); OpenAI’s Image prompting guide (13 Oct 2026) and Image generation guide (16 Sep 2026); Midjourney’s No parameter page (16 Sep 2026); Runway’s Gen-4 Video Prompting Guide (17 Sep 2026); Kling’s Text-to-video prompt guide (16 Sep 2026).
All 30 prompts and the “change this” lines are this site’s own examples, not claims from these sources.
Free printable AI cheat sheets, and AI news every week.
Sixteen one-page AI guides and checklists in one free PDF, no sign-up. And if you want to keep up, the newsletter covers what changed in AI each week, in plain English, with the sources.
SUBSCRIBE FREE ↗ GET THE CHEAT SHEET PACK