# AI use cases for Yap (dzhng/yap) Repo: [dzhng/yap](https://github.com/dzhng/yap) | Use cases: 5 | Page: https://dedupedai.com/use-cases/dzhng-yap/ Yap is a macOS app and agent toolkit that records your screen, voice and pointer marks so AI agents like Claude Code and Codex can read, edit and act on the recordings. This file holds every use case of this repo. Each one starts with a "## " heading and ends with its URL. - Show your agent a bug and have it fixed: https://dedupedai.com/use-cases/dzhng-yap/show-agent-bug-fix/ - Turn a rough product take into a captioned demo: https://dedupedai.com/use-cases/dzhng-yap/turn-rough-take-into-demo/ - Clean a podcast episode into clips and show notes: https://dedupedai.com/use-cases/dzhng-yap/clean-podcast-into-clips-and-notes/ - Rehearse a pitch on camera and get pacing feedback: https://dedupedai.com/use-cases/dzhng-yap/rehearse-pitch-on-camera/ - Fix a misspoken word in your own voice: https://dedupedai.com/use-cases/dzhng-yap/fix-misspoken-word-in-own-voice/ ## How to show your agent a bug and have it fixed with Yap Repo: [dzhng/yap](https://github.com/dzhng/yap) | Difficulty: Beginner | Time: 15 minutes | Basis: From the README Record the bug on screen, pin the broken element, then paste the prompt so your agent can fix it. **Who it is for:** A developer or designer who would rather show a visual bug than write three paragraphs describing it to Claude Code or Codex. **The problem:** Describing a visual bug or design change in text takes long and still leaves the agent guessing which element you meant. **The outcome:** The agent receives the recording, your spoken words, your pins and the time of each mark, and makes the change you pointed at. ### Steps 1. Ask your agent to install Yap by following https://justyap.io/llms.txt, or download the latest release from GitHub Releases and open Yap.app. 2. Press ⌥⇧S over the display with the bug, talk while you hold ⇧ to draw, and ⇧-click to drop a numbered pin on the broken element. 3. Press ⌥⇧S again to finish the recording. 4. Press ⌥⇧V to paste the prompt into your agent, then say 'fix this'. 5. Start a fresh agent session so the Yap skill is discovered, using $Yap for Codex or /yap for Claude Code. (inferred, check the README) ### Prompt to paste into your agent ```text Using the Yap CLI and skill in the , read the recording referenced in the pasted Yap prompt, find the frame where I said 'this' and the element I pinned, and fix it in the code. Report the changed files and confirm the pinned element now renders correctly. ``` ### Caveat Releases are not notarized, so a browser download needs Open Anyway in System Settings before Yap will run. Risk, remote script: The install prompt tells your agent to follow instructions from a web page, so check the SHA256SUMS checksum of the downloaded ZIP before extracting it. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/dzhng-yap/show-agent-bug-fix/ ## How to turn a rough product take into a captioned demo with Yap Repo: [dzhng/yap](https://github.com/dzhng/yap) | Difficulty: Intermediate | Time: 30 minutes | Basis: From the README Have your agent cut a single take of you showing the product into a tight, captioned demo video. **Who it is for:** A founder or marketer preparing a launch who has one unpolished screen recording and needs a demo video today. **The problem:** A single rough take has dead air, false starts and no captions, and editing it by hand takes hours. **The outcome:** A tighter MP4 with captions seeded from the transcript, cut to the length you asked for, plus a non-destructive revision you can undo. ### Steps 1. Record the product walkthrough with ⌥⇧S, or hand Yap a video you already have. 2. Wait for the automatic transcription to finish before asking for edits. 3. Tell your agent the length and style you want, such as 'a tight 60-second product demo with captions'. 4. Ask the agent to review the result and export the final video. 5. Check the exported file and ask for another revision if a cut is wrong. ### Prompt to paste into your agent ```text Using the Yap CLI and skill, take the recording in , cut out dead air and false starts, add captions from the transcript, and export a 60-second MP4 demo. Report the output path and the total duration, and confirm the captions match the spoken words. ``` ### Caveat Yap supplies the editing primitives but does not decide what to cut, so the quality of the demo depends on the instructions your agent receives. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/dzhng-yap/turn-rough-take-into-demo/ ## How to clean a podcast episode into clips and show notes with Yap Repo: [dzhng/yap](https://github.com/dzhng/yap) | Difficulty: Intermediate | Time: 30 minutes | Basis: From the README Clean the audio, cut highlight clips with captions and turn the transcript into show notes. **Who it is for:** A podcaster or talk host who recorded an episode and wants clips and written notes without spending a full day editing. **The problem:** Podcast audio needs cleanup, levels that match, short captioned clips and written notes, and each step is slow by hand. **The outcome:** Clean audio with consistent loudness, captioned highlight clips, and a transcript your agent can turn into show notes or a post. ### Steps 1. Import the episode into Yap as a video you already have, or record it with ⌥⇧S. 2. Wait for the automatic transcription to finish. 3. Ask your agent for clean narration audio, loudness matched across the episode, and a noise-reduced M4A export. 4. Ask for three highlight clips with captions, each exported as an MP4. 5. Ask for show notes built from the transcript and check them against the audio before publishing. ### Prompt to paste into your agent ```text Using the Yap CLI and skill, process the episode at . Remove background noise, match loudness, export an M4A of the narration, and cut three captioned highlight clips. Report each output path and check that the show notes quote the transcript accurately. ``` ### Caveat Yap does not choose which moments are highlights, so your agent and your own review decide the clip selection. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/dzhng-yap/clean-podcast-into-clips-and-notes/ ## How to rehearse a pitch on camera and get pacing feedback with Yap Repo: [dzhng/yap](https://github.com/dzhng/yap) | Difficulty: Intermediate | Time: 15 minutes | Basis: From the README Record a practice pitch on camera and have your agent report pacing, filler words and rambling sections. **Who it is for:** A founder or speaker preparing for a demo day, investor meeting or conference talk who wants evidence about how they delivered the material. **The problem:** Practising in front of a mirror gives no measurable feedback on pacing, filler words or where the talk drifts. **The outcome:** A report based on the transcript and audio that points to the sections where you rambled and the words you overused. ### Steps 1. Press ⌥⇧S to record the practice pitch, with the camera enabled in the menu bar recording options. 2. Press ⌥⇧S again to finish the recording. 3. Wait for the automatic transcription to complete. 4. Ask your agent how your pacing was and where you rambled, and which filler words came up most. 5. Keep your best take if you want, and ask the agent to keep only that take. ### Prompt to paste into your agent ```text Using the Yap CLI and skill, read the transcript and audio of the practice recording . Report the pacing, the filler words with their timestamps and any sections where the pitch drifted. Verify each timestamp by checking the transcript word timings. ``` ### Caveat The feedback comes from the transcript and audio measurements, so it describes what was said and how it sounded rather than how it looked on camera. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/dzhng-yap/rehearse-pitch-on-camera/ ## How to fix a misspoken word in your own voice with Yap Repo: [dzhng/yap](https://github.com/dzhng/yap) | Difficulty: Advanced | Time: 30 minutes | Basis: From the README Regenerate one misspoken phrase in your voice with a local model instead of re-recording the whole take. **Who it is for:** A creator or presenter who misspoke one word in an otherwise good recording and does not want to re-record it. **The problem:** Re-recording a whole take to fix one word wastes time and can make the rest of the take harder to match. **The outcome:** The misspoken phrase is replaced in your own voice, and the edit is saved as a revision you can undo. ### Steps 1. Record the take with ⌥⇧S, or import a video you already have. 2. Wait for the automatic transcription, then find the phrase that was misspoken. 3. Tell your agent the correct word, for example 'I said version three at 0:42 but meant version two'. 4. Ask the agent to regenerate that phrase in your voice with the local model. 5. Preview the revision and undo it if the replacement does not sound right. ### Prompt to paste into your agent ```text Using the Yap CLI and skill, find the phrase at in recording and regenerate it in the speaker's own voice, replacing the word with . Report the revision ID and confirm by previewing that only that phrase changed. ``` ### Caveat The voice model runs locally on your Mac, and the source recording stays byte-for-byte as you captured it. Risk, voice cloning: This use case generates speech in the speaker's own voice, so only do it with recordings of your own voice or with clear permission from the speaker. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/dzhng-yap/fix-misspoken-word-in-own-voice/