AI skill marketplace
Find the skill that fits the work you actually do, and widen what your AI can take on. No setup, no prompt writing.
This only searches. Nothing runs and nothing is charged until you pick a skill yourself.
Skills for sale
Paste the job posting and your current résumé, and get back a version written for that one posting: the same facts, reordered and reworded around what the employer actually asked for, with the gaps named so you know what you are walking into rather than finding out in the interview.
영상과 시청자를 설명하면 크리에이티브 방향을 받습니다. 정확한 색상과 글꼴을 담은 frame.md 디자인 명세, 장면별 비트 계획, 장면 길이에 맞춘 화면 문구와 내레이션, 모든 컴포지션에 붙여 넣을 CSS 토큰까지 들어 있어 영상 전체가 하나로 보입니다.
Attach an SRT or VTT file and name a language, and get back the translated file ready to load: every cue and timecode untouched, lines kept short enough to read at the speed the subtitles pass, and phrasing condensed where a literal translation would not fit on screen.
Attach a CSV and say what the video should show, and get back a Remotion project that turns your numbers into an animated chart video: bars that grow, lines that draw, values that count up, and one takeaway sentence taken from the data rather than made up.
Attach a report, a proposal or a long memo and get back the deck it should be: a slide at a time, with the line that goes on the slide, the few words underneath it, and what you would say out loud. Built around the argument rather than around the section headings.
Attach a screenshot, a photo of a sign, a whiteboard or a page of a book, and get the text back as text you can select, search and paste. Tables come back as tables, and anything too blurred to be sure of is marked rather than guessed at.
HyperFrames 영상 프로젝트로 하려는 일(새로 시작하기, 미리보기, 실패하는 검사 통과시키기, 최종 MP4 렌더링)을 적거나 마주친 오류를 붙여 넣으면, 실행할 명령어를 순서대로, 각 단계가 정상일 때 나오는 결과와 그렇지 않을 때의 대처법과 함께 알려 드립니다.
Describe the camera move you want on a photo, a screenshot or a clip (a slow push-in, a punch-in on a detail, a pan, a Ken Burns drift, a whip into the next shot) and get a HyperFrames scene with the move keyframed seek-safe on a wrapper, so it renders exactly as it previews.
Attach a video's transcript and get back what goes around it on a video platform: chapters with timestamps where the topic really changes, a description whose first two lines carry the promise, five honest title options, search tags and a pinned comment.
Tell it your topic, product or story, and get back a script for one Reel, Short or TikTok: five hook options for the first two seconds, a shot-by-shot table of what is said and shown, on-screen text, and a caption with hashtags, with no invented claims.
Attach a meeting transcript and get back the three things anybody actually needs from it: what was decided, who owes what by when, and what was left open. Anything that was only discussed and never settled is listed as open rather than written up as a decision.
Attach a PDF or a Word file and get back what is actually in it: the points that matter, the things it is asking you to do, the dates it commits you to, and the parts vague enough to be worth asking about. Written for a document you have to act on rather than one you are studying.
Describe a short video (what it says, how long, where it will be posted) and get back a complete HyperFrames project: one HTML composition per scene, ready to check, preview and render to MP4 on your own computer, with every word, colour and timing easy to change.
영상에 필요한 효과나 분위기(글리치 전환, 필름 그레인, 쌓여 올라가는 차트, 자막, 코드 화면, 하단 자막 바)를 설명하면, 이미 그 효과를 구현한 레지스트리 블록과 컴포넌트, 검색하고 설치하는 명령어, 프로젝트에 연결하는 정확한 마크업을 알려 드립니다.
Describe a maths or science idea and get back a Manim scene that renders on an ordinary computer: one Python file and one command, with no LaTeX to install and no Python setup to fight, animating the idea step by step the way a good teacher would explain it.
Attach your subtitle file and get back a Remotion project that plays your own video with bold word-by-word captions on top, the style short-form feeds use: the spoken word highlighted, large outlined text that reads on any footage, and positioned clear of the app's buttons.
Attach a CSV or paste a table and get back the few things in it worth knowing: what stands out, what is moving, and where the data is thin enough that you should not lean on it. Written in sentences, with the figures that back each one, for the person who has to explain the sheet rather than build it.
Attach a document and name the language you want it in. Headings stay headings, lists stay lists, and the numbers, names and dates come through untouched. Anything that should not have been translated — a product code, a legal term of art, a company name — is left alone and listed at the end so you can check.
짧은 영상의 내용과 길이, 올릴 곳을 적으면 완성된 HyperFrames 프로젝트를 받습니다. 장면마다 HTML 컴포지션이 하나씩 들어 있고, 내 컴퓨터에서 바로 검사하고 미리 보고 MP4로 렌더링할 수 있으며, 문구와 색상, 타이밍을 쉽게 바꿀 수 있습니다.
렌더링이 안 되거나, 빈 화면 또는 멈춘 화면으로 나오거나, 깜박이거나, 검사를 통과하지 못하는 HyperFrames 컴포지션을 첨부하거나 붙여 넣으면, 모든 문제를 원인이 된 줄과 어긴 규칙과 함께 짚고, 바로 다시 검사할 수 있게 고친 파일 전체를 드립니다.
Describe an animation — an explainer, a diagram that builds, a title sequence — and get back a complete Motion Canvas project that installs and starts on the first try, with pinned versions and the configuration fixes the official setup leaves you to discover.
Attach the transcript of a video, lecture or podcast you learned something from and get back a plan you can start this week: the core idea, three things to do in the next 48 hours, a four-week plan, the key techniques with timestamps, and a printable checklist.
Attach something written for the inside — a policy, a release note, an engineer's explanation of an outage — and get it back in words a customer can read. Same facts, no internal names, nothing softened into vagueness, and a list of anything you should check before it goes out.
Attach a contract and find out which parts work against you before you sign. Every clause that puts the risk, the cost or the exit on your side is quoted, explained in ordinary words, and given a sentence you could actually send back. It is a careful reading rather than legal advice, and it says so.
사진, 스크린숏, 영상 클립에 원하는 카메라 움직임(천천히 다가가기, 디테일로 펀치 인, 패닝, 켄 번스 효과, 다음 장면으로의 휩)을 설명하면, 그 움직임을 래퍼에 시크 안전한 키프레임으로 걸어 둔 HyperFrames 장면을 받습니다. 미리보기와 똑같이 렌더링됩니다.
Remotion 컴포넌트나 프로젝트 파일을 첨부하거나 붙여 넣으면 Remotion의 렌더링 규칙(프레임 기반 움직임, 타이밍 계산, 글꼴, 미디어, 시퀀스) 기준으로 검토해, 모든 문제를 쉬운 말로 설명하고 고친 코드 전체를 드립니다. 미리보기와 똑같이 렌더링되게 만듭니다.
Describe your product and customer and get back five scripts for creator-style video ads, each from a different angle — problem and solution, before and after, objection-busting and more — written the way people actually talk to their phone, using only claims you can back up.
Paste the script or a shot-by-shot description of a video ad — a competitor's or your own — and get back why it works or does not: the hook, the structure, the angle, the proof and the offer, researched on the web, plus three techniques to adapt without copying.
Attach an invoice or a receipt — a PDF or a photograph of the paper — and get the line items back as rows you can paste straight into a spreadsheet. Totals are checked against the lines, and anything the image was too blurred to read is marked rather than guessed.
대본이나 장면 목록을 붙여 넣으면 HyperFrames 영상의 미디어 계획을 받습니다. 장면마다 필요한 배경음악, 효과음, 이미지, 아이콘, 로고, 내레이션과 이를 찾거나 만드는 media-use 명령어, 장면에 맞춘 내레이션 원고, 각 파일을 배치하는 마크업까지 들어 있습니다.
Attach or paste a Remotion component or project file and get it reviewed against the rules Remotion renders by (frame-driven motion, timing maths, fonts, media, sequencing), with every problem explained in plain words and the corrected code in full, so the video renders the way it previews.
Attach the timestamped transcript of a long video or podcast and get back the five moments most likely to work as standalone Shorts or Reels, with exact start and end times, the hook line, a title and what to trim, so you edit the right clips first.
Attach a research paper as a PDF and get back the plan for a short explainer video about it: a one-sentence version for non-specialists, a scene-by-scene script with what is on screen, a slide outline, and the claims the authors should double-check before publishing.
장면과 그 움직임(자리 잡는 제목, 하나씩 쌓이는 목록, 올라가는 숫자, 다음 장면으로의 전환)을 설명하면, 움직임을 시크(seek)에 안전한 GSAP 타임라인 하나로 박자마다 맞춰 쓴 HyperFrames 장면을 받아 프로젝트에 바로 넣을 수 있습니다.
Say what you want to do with a HyperFrames video project (start one, preview it, get a failing check to pass, render the final MP4) or paste the error you hit, and get the exact commands to run in order, what each should print, and what to do when it does not.
Describe what the narration is for and how long it must run, and get back a script written to be spoken — for your own voice, a voice actor, text-to-speech or an AI avatar — timed to the second, with pauses and emphasis marked and a clean copy for tools that do not read marks.
Attach a short clip and describe one change (a new background, a different style, an outfit or product swap) and get the first five seconds back edited by Wan 2.7, with the subject, their movement and the camera kept as they were and the original sound left in.
짧은 클립을 첨부하고 바꿀 점 하나(새 배경, 다른 스타일, 옷이나 제품 교체)를 적으면, Wan 2.7이 편집한 앞부분 5초를 받습니다. 인물과 움직임, 카메라 움직임은 그대로 두고 원래 소리도 유지합니다.
Describe your video and its audience and get its creative direction: a frame.md design spec with exact colours and fonts, a beat plan scene by scene, on-screen copy and narration paced to each scene, and the CSS tokens to paste into every composition so the whole video reads as one piece.
Paste your script or scene list and get the media plan for a HyperFrames video: the music bed, sound effects, images, icons, logos and voiceover each scene needs, the media-use commands that fetch or generate them, the narration timed to the scenes, and the markup that places every file.
Attach or paste a HyperFrames composition that will not render, renders blank or frozen, flickers or fails its check, and get back every problem tied to the line that causes it and the rule it breaks, with the corrected files in full, ready to check again.
Describe the look or effect your video needs (a glitch transition, film grain, a chart that builds, captions, a code window, a lower third) and get the registry blocks and components that already do it, the commands to search for and install them, and the exact markup to wire each one in.
Describe a scene and how it should move (a title that lands, a list that builds, a number that counts up, a transition into the next shot) and get a HyperFrames scene with the motion written as one seek-safe GSAP timeline, timed beat by beat, ready to drop into your project.
Turn a topic, draft, or product idea into a YouTube script with a strong hook, clear sections, visual cues, and a closing call to action. Supports Shorts and long-form videos.
Tell it what your video is about, and optionally the words you want on it, and get back a 16:9 thumbnail built to be read at the size of a phone search result: one big subject, strong contrast, and at most five words in heavy letters spelled exactly as you gave them.
Write the story of a short video in a few sentences and get back one storyboard sheet: six numbered panels in order, the same characters and places from panel to panel, shots varied the way a director would, and a short caption under each. The plan you shoot or generate from.
Describe a scene in a sentence or two and get back a five-second vertical clip made for Reels, Shorts and TikTok: one continuous shot, one deliberate camera move, and the action kept in the middle of the frame where the app's buttons and captions will not cover it.
Upload one product photo and get back a five-second ad shot built from it: the camera moves, the product does not change. Same shape, same label, same colours, with slow commercial camera work and soft light moving across it, ready to drop into an ad or a product page.
Upload a character, mascot or logo and get back a short animation of it doing one simple gesture — a wave, a nod, a bounce — that ends in the same pose it started in, so it loops cleanly as a reaction, a sticker or a story overlay.
Describe the video — what it says, how long, where it is going — and get back a complete Remotion project that renders on the first try: every file in full, the exact setup commands, and the words and colours kept in one place so you can change them without reading the code.
Send the photograph you took on your phone and get back the one a store page needs: the product lit evenly, the clutter gone, the background clean and consistent across everything you sell. Built for people who have a hundred things to photograph and no studio.
Describe the picture you want and it draws one — no photograph needed, nothing to upload. Built for a feed rather than a gallery: one subject, high contrast, composed so it still reads at the size a thumb actually sees it, with room left for you to lay a headline over the top.
Name a topic and it looks the subject up before writing a word, comes back with the angle and the outline it intends, and waits for you to approve or redirect. Then it writes the piece against current sources with the links kept, so you can check what it used rather than trusting it.
Send one photograph of the thing you sell and get back a few seconds of video: a slow move around it, the light shifting the way it would if someone picked it up. Made for a product page or a feed, where a still gets scrolled past and a moving thing does not.
For sellers
Write the instructions for your task, select a supported model, and test the result. Saved instructions stay out of the buyer interface.
Charge by the hour, by the day, by the month, or once and done. Offer more than one price on the same skill if that fits the work.
After review, it appears in the category you pick. Creator earnings are credited when payment is confirmed, with refunds and disputes reflected in the balance.
How it works
Coverage
Code, research, video, books, tickets, spreadsheets. Every skill here was built for one kind of work by a person who does that work.
Access
Open a skill, describe the task, and it runs on our machines. No account with a model provider, no API key, no setup on your side.
Price
By the hour for one-off work, by the week for a push, monthly when it becomes routine, or once and done. Cancel any of them the same day.
Privacy
The model provider processes your input through ESF. Creators see usage summaries, without access to another buyer's raw brief or results. Saved instructions are not exposed through the buyer interface.
Questions
Still unclear on something? Write to desk@everyskillfinder.com and a person answers.
Instructions someone wrote so an AI model does one job well — reworking a résumé for a particular posting, cutting an interview into shorts, checking a contract. Somebody who does that work for a living wrote them, tested them, and put them up for sale.
Describe your task in the box at the top, or browse by category. Open the skill that fits, tell it what you need, and the finished result comes back. It runs on our machines, in your browser — there is nothing to install and no API key to get.
We do not sell it and we do not train models on it. It is encrypted in transit and at rest. Skills are built by independent sellers, and a skill may pass your input to its own systems to finish the job — each listing says what it touches before you start.
The seller sets the price: by the hour, by the day or week, monthly, or once and done. One skill can offer more than one. You see the price before you pay, and a subscription is cancelled from your own page.
Yes, on the same account you buy with. Package what you already use, set the price, and it goes up for sale. You keep 85% of every sale and withdraw it to your bank whenever you like.
You pay for access to a skill, not per run, so a failed run costs nothing on its own — it does not even spend one of your runs. If a skill takes your money and does not do what it says, write to support@everyskillfinder.com within seven days and we will sort it out.
