# AI use cases for Axuse (ctoldigital/axuse) Repo: [ctoldigital/axuse](https://github.com/ctoldigital/axuse) | Use cases: 4 | Page: https://dedupedai.com/use-cases/ctoldigital-axuse/ axuse is a zero-dependency MCP server that lets AI agents drive a Chromium browser (and, on Windows, the desktop) through the accessibility tree and CDP instead of screenshot loops. This file holds every use case of this repo. Each one starts with a "## " heading and ends with its URL. - Find a fact on a web page and confirm its source: https://dedupedai.com/use-cases/ctoldigital-axuse/verify-fact-on-web-page/ - Fill a long web form in one call and submit it: https://dedupedai.com/use-cases/ctoldigital-axuse/fill-long-form-in-one-call/ - Pull matching lines from a long or endless web page: https://dedupedai.com/use-cases/ctoldigital-axuse/extract-matching-lines-from-long-page/ - Let an agent work inside your own logged-in Chrome: https://dedupedai.com/use-cases/ctoldigital-axuse/work-in-logged-in-daily-chrome/ ## How to find a fact on a web page and confirm its source with Axuse Repo: [ctoldigital/axuse](https://github.com/ctoldigital/axuse) | Difficulty: Intermediate | Time: 30 minutes | Basis: From the README An agent opens a page, locates a named link or value, and reports where it points. **Who it is for:** A journalist or researcher who needs to check a claim on a specific web page before publishing or citing it. **The problem:** Screenshot-based agents guess pixel positions and often misclick or miss small text, so the answer is slow to get and hard to trust. **The outcome:** The agent returns the matching element text and its destination, based on the accessibility snapshot rather than an image guess. ### Steps 1. `git clone https://github.com/ctoldigital/axuse.git` 2. `cd axuse` 3. `node src/cli.js doctor` 4. `codex mcp add axuse -- node /abs/path/axuse/src/cli.js mcp` 5. Ask the agent to open the target URL, find the link, and report where it goes. ### Prompt to paste into your agent ```text Using the axuse MCP server from the repo at , open , find the link labeled , and report its destination. Check that the reported destination matches the link target shown in the accessibility snapshot before answering. ``` ### Caveat Axuse needs a local Node.js 22 install and a clone of the repo, so it is not a copy-and-paste setup for non-developers; use the non-developer onboarding guide in the repo for that path. Dual-use notice: only use this on software, sites and systems you own or are authorized to test. 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/ctoldigital-axuse/verify-fact-on-web-page/ ## How to fill a long web form in one call and submit it with Axuse Repo: [ctoldigital/axuse](https://github.com/ctoldigital/axuse) | Difficulty: Intermediate | Time: 30 minutes | Basis: From the README An agent sets every field of a multi-step web form and clicks submit, with typed errors on failure. **Who it is for:** An operations or admin person who must enter the same details into a long web form for many records. **The problem:** Filling forms field by field takes many steps, and a single misclick can silently enter the wrong value. **The outcome:** The form is filled in one tool call and the agent gets a typed error with a retry hint if a field fails. ### Steps 1. `git clone https://github.com/ctoldigital/axuse.git` 2. `cd axuse` 3. `node src/cli.js print-config --harness claude-code` (inferred, check the README) 4. `claude mcp add --scope user axuse -- node /abs/path/axuse/src/cli.js mcp` 5. Give the agent the form URL and the values, and ask it to confirm the values on screen before it clicks submit. ### Prompt to paste into your agent ```text Using the axuse MCP server, open
and fill the form with these values: . Use one fill action, then read the accessibility snapshot to confirm each field holds the intended value before submitting. Report any field that returns an error code. ``` ### Caveat The README says ax_fill can click submit, so the agent should be told to check values before submitting rather than assuming the form is correct. Dual-use notice: only use this on software, sites and systems you own or are authorized to test. 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/ctoldigital-axuse/fill-long-form-in-one-call/ ## How to pull matching lines from a long or endless web page with Axuse Repo: [ctoldigital/axuse](https://github.com/ctoldigital/axuse) | Difficulty: Intermediate | Time: 30 minutes | Basis: From the README An agent scrolls a feed until a term appears, then returns only the matching lines. **Who it is for:** A researcher or analyst who needs to find mentions of a term in a long feed, infinite list, or many pages. **The problem:** Loading a long page into an agent's context uses many tokens, and infinite lists require manual scrolling. **The outcome:** The agent returns only the lines that match the term, with the page text it found, instead of the whole page. ### Steps 1. `git clone https://github.com/ctoldigital/axuse.git` 2. `cd axuse` 3. `node src/cli.js doctor` 4. `codex mcp add axuse -- node /abs/path/axuse/src/cli.js mcp` 5. Ask the agent to scroll until the term appears, then observe only the matching lines. ### Prompt to paste into your agent ```text Using the axuse MCP server, open , scroll until the text appears, then observe only the lines matching that text. Return the matching lines as a list, and check that each line appears in the page when you observe it again. ``` ### Caveat The README describes ax_open_many as a bounded list, so very large batches must be split into several runs. Dual-use notice: only use this on software, sites and systems you own or are authorized to test. 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/ctoldigital-axuse/extract-matching-lines-from-long-page/ ## How to let an agent work inside your own logged-in Chrome with Axuse Repo: [ctoldigital/axuse](https://github.com/ctoldigital/axuse) | Difficulty: Advanced | Time: 45 minutes | Basis: From the README An agent uses your daily Chrome or Edge session through the extension bridge. **Who it is for:** A knowledge worker who needs an agent to act inside sites they are already logged into, without signing in again. **The problem:** A fresh browser has no logins, so agents cannot reach dashboards or accounts the user already uses. **The outcome:** The agent runs actions in the user's existing browser session after a one-time extension setup and a verified connection. ### Steps 1. `git clone https://github.com/ctoldigital/axuse.git` 2. `cd axuse` 3. `node src/cli.js setup --browser chrome --harness codex` 4. Click Load unpacked once on the extensions page that opens, then run `node src/cli.js bridge verify`. 5. Ask the agent to open the site with the extension backend and report what it sees on the logged-in page. ### Prompt to paste into your agent ```text Using the axuse MCP server with the extension backend, open in the user's logged-in Chrome session and report the account name or main heading shown on the page. Confirm the bridge is connected with the bridge verify command before acting. ``` ### Caveat The README says the extension bridge drives the user's real browser, so the agent sees and changes whatever that session is already logged into. Risk, credentials: The agent acts inside your logged-in sessions, so it can reach any account you are signed into; limit tasks to ones you have approved and review actions as they happen. Dual-use notice: only use this on software, sites and systems you own or are authorized to test. 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/ctoldigital-axuse/work-in-logged-in-daily-chrome/