# AI use cases for Jevry (michaelswissa/jevry) Repo: [michaelswissa/jevry](https://github.com/michaelswissa/jevry) | Use cases: 4 | Page: https://dedupedai.com/use-cases/michaelswissa-jevry/ Jevry is an MIT-licensed desktop browser agent (Electron, TypeScript) that carries out website tasks in plain language and reports results grounded in the visible page. This file holds every use case of this repo. Each one starts with a "## " heading and ends with its URL. - Compare two web pages and cite each claim: https://dedupedai.com/use-cases/michaelswissa-jevry/compare-sources-with-citations/ - Find a library's getting-started guide in docs: https://dedupedai.com/use-cases/michaelswissa-jevry/find-library-getting-started-guide/ - Collect records that match filters in a web table: https://dedupedai.com/use-cases/michaelswissa-jevry/collect-matching-catalog-records/ - Have Jevry play and win a 2048 game: https://dedupedai.com/use-cases/michaelswissa-jevry/play-numeric-puzzle-game/ ## How to compare two web pages and cite each claim with Jevry Repo: [michaelswissa/jevry](https://github.com/michaelswissa/jevry) | Difficulty: Intermediate | Time: 60 minutes | Basis: From the README Ask Jevry to read two or more pages and return a comparison with citations to the pages it read. **Who it is for:** A researcher or analyst who must compare policies, terms, or specs across several sites and needs cited answers quickly. **The problem:** Reading several pages side by side and copying exact wording into a summary is slow and easy to get wrong. **The outcome:** A comparison answer with citations to the source pages, which you can check against the visible browser tabs. ### Steps 1. Run `git clone https://github.com/michaelswissa/jevry.git` and then `cd jevry`. 2. Run `npm ci`, then run `npm run dev` to open the Electron desktop app. 3. Enter your TypeSafe Jev API key and connect a text model in the app setup. 4. Open the first source page, then ask Jevry to compare it with the second page and cite both. ### Prompt to paste into your agent ```text Use the repo michaelswissa/jevry: after starting it with npm run dev, ask the desktop app to compare the cancellation policies at and and cite each claim. Expect a comparison with citations; verify each cited sentence appears on the matching visible page. ``` ### Caveat The README describes research and citation as a live fixed-corpus study, so check every citation yourself before relying on 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/michaelswissa-jevry/compare-sources-with-citations/ ## How to find a library's getting-started guide in docs with Jevry Repo: [michaelswissa/jevry](https://github.com/michaelswissa/jevry) | Difficulty: Intermediate | Time: 45 minutes | Basis: From the README Point Jevry at a project's public documentation site and have it navigate to the getting-started page. **Who it is for:** A developer evaluating an unfamiliar library who wants the setup guide without clicking through a documentation site by hand. **The problem:** Documentation sites have nested menus and search boxes, and finding the right starting page takes several clicks. **The outcome:** The browser lands on the getting-started guide, with each action visible in the action feed. ### Steps 1. Run `git clone https://github.com/michaelswissa/jevry.git` and then `cd jevry`. 2. Run `npm ci`, then run `npm run dev`. 3. Connect your TypeSafe Jev key and a text model in setup. 4. Open the library's public docs homepage and ask Jevry to find its getting-started guide; press Stop or Escape to cancel if needed. ### Prompt to paste into your agent ```text Use the repo michaelswissa/jevry: run npm run dev, then ask the desktop app to find the getting-started guide on . Expect the browser to end on that guide page; verify the page title and heading match what you asked for. ``` ### Caveat The README gives no success rate for documentation navigation, so test it on the specific site you need before trusting 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/michaelswissa-jevry/find-library-getting-started-guide/ ## How to collect records that match filters in a web table with Jevry Repo: [michaelswissa/jevry](https://github.com/michaelswissa/jevry) | Difficulty: Advanced | Time: 60 minutes | Basis: From the README Have Jevry apply filters to a catalog or admin table and report the matching records it found. **Who it is for:** An operations person who must pull a list of matching records from a web admin or catalog page each week. **The problem:** Applying filters and reading results across many pages by hand is repetitive and invites missed rows. **The outcome:** A list of matching records from the table, with the caveat that you must confirm every page was covered. ### Steps 1. Run `git clone https://github.com/michaelswissa/jevry.git` and then `cd jevry`. 2. Run `npm ci`, then run `npm run dev`. 3. Connect your TypeSafe Jev key and a text model in setup. 4. Open the table page, ask Jevry to apply your filters and list matching records, then count the pages it reports reading against the pager. ### Prompt to paste into your agent ```text Use the repo michaelswissa/jevry: after npm run dev, ask the desktop app to find records on matching . Expect a list of records; verify the count against the table's own total and confirm every result page was read. ``` ### Caveat The README lists multi-page completeness as a known failure mode, so a short result list does not prove the table was fully read. Risk, credentials: Pages behind a login expose private records, so use only accounts you are authorized to use and keep the page visible while Jevry works. 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/michaelswissa-jevry/collect-matching-catalog-records/ ## How to have Jevry play and win a 2048 game with Jevry Repo: [michaelswissa/jevry](https://github.com/michaelswissa/jevry) | Difficulty: Intermediate | Time: 45 minutes | Basis: From the README Ask Jevry to play a supported numeric puzzle game in the browser and report the final board state. **Who it is for:** A hobbyist who wants to watch an agent play a browser game and check its moves on screen. **The problem:** Playing a puzzle game through to a win takes time, and watching an agent solve it is a useful test of how it handles page state. **The outcome:** The game reaches a winning tile, with the score and move count reported from the visible page. ### Steps 1. Run `git clone https://github.com/michaelswissa/jevry.git` and then `cd jevry`. 2. Run `npm ci`, then run `npm run dev`. 3. Connect your TypeSafe Jev key and select a text-provider and model that supports image input. 4. Open a 2048 page and ask Jevry to play and win it, then watch the action feed until the game reports a result. ### Prompt to paste into your agent ```text Use the repo michaelswissa/jevry: run npm run dev, then ask the desktop app to play and win the 2048 game at . Expect a win message with a score and move count; verify the win state appears on the visible game board. ``` ### Caveat The README's 2048 victory is a recorded run from an earlier interface, so a new run may not match it and the game may need a vision-capable model. 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/michaelswissa-jevry/play-numeric-puzzle-game/