const env = await trycase.launch({ os: "linux" });
await env.upload(repo);
await env.exec("bun install && bun dev");

const browser = await env.browser.open(appUrl);
await browser.click("[data-test=checkout]");
await browser.fill("#email", user.email);

const proof = await env.record({ video: true });
await browser.expect(page).toShow("Order confirmed");

return {
  screenshots: proof.screenshots,
  recording: proof.video,
  logs: env.logs.tail(200),
};

// iterate until the flow passes
while (!result.passed) {
  await agent.fix(result.failure);
  result = await env.retest();
}

const env = await trycase.launch({ os: "linux" });
await env.upload(repo);
await env.exec("bun install && bun dev");

const browser = await env.browser.open(appUrl);
await browser.click("[data-test=checkout]");
await browser.fill("#email", user.email);

const proof = await env.record({ video: true });
await browser.expect(page).toShow("Order confirmed");

return {
  screenshots: proof.screenshots,
  recording: proof.video,
  logs: env.logs.tail(200),
};

// iterate until the flow passes
while (!result.passed) {
  await agent.fix(result.failure);
  result = await env.retest();
}

  0 00000000 0000000000000 0  0000
00000000000000000000  0  000000000
000 0000 00  00100000000 00000 00 
0000  000 0 0 0  00000000000000 00
0000  0000000000000 00  0000000000
0000000  00 00 00000  000000000000
00000 0 00 0 00000 00000  000 000 
00  0000000 000000000 000000000   
00 00000000000 0000000 000000 00 0
000   00000 00000000000    00 000 
10000 0000000000 0 000000000  0000
0000000 00010000 000000000 00 000 
0000000000000000  000000000000 000
000 000 0 000 000000 000000 000000
  0000 000000000 000 0 00000000 00
000100 0 0  0000 00000 0000 00000 
 0  00000000000 000000    00000000
0 0 0000 000  00   00 0 00 00 0 00
0000000000000 0 0000000000 0 0 00 
000000000  00    0 000000000 000  
00 0000 000 0 00000000 000000 0000
0  00 0 0000000 000 00 0000 00000 
000000 000  000 000000 00000000000
0 00 0000000000 0 00000000  00000 
 0000000 000000 00 000 00000000000
0000 000000000 000000 0000 000 000

const env = await trycase.launch({ os: "linux" });
await env.upload(repo);
await env.exec("bun install && bun dev");

const browser = await env.browser.open(appUrl);
await browser.click("[data-test=checkout]");
await browser.fill("#email", user.email);

const proof = await env.record({ video: true });
await browser.expect(page).toShow("Order confirmed");

return {
  screenshots: proof.screenshots,
  recording: proof.video,
  logs: env.logs.tail(200),
};

// iterate until the flow passes
while (!result.passed) {
  await agent.fix(result.failure);
  result = await env.retest();
}

const env = await trycase.launch({ os: "linux" });
await env.upload(repo);
await env.exec("bun install && bun dev");

const browser = await env.browser.open(appUrl);
await browser.click("[data-test=checkout]");
await browser.fill("#email", user.email);

const proof = await env.record({ video: true });
await browser.expect(page).toShow("Order confirmed");

return {
  screenshots: proof.screenshots,
  recording: proof.video,
  logs: env.logs.tail(200),
};

// iterate until the flow passes
while (!result.passed) {
  await agent.fix(result.failure);
  result = await env.retest();
}

const env = await trycase.launch({ os: "linux" });
await env.upload(repo);
await env.exec("bun install && bun dev");

const browser = await env.browser.open(appUrl);
await browser.click("[data-test=checkout]");
await browser.fill("#email", user.email);

const proof = await env.record({ video: true });
await browser.expect(page).toShow("Order confirmed");

return {
  screenshots: proof.screenshots,
  recording: proof.video,
  logs: env.logs.tail(200),
};

// iterate until the flow passes
while (!result.passed) {
  await agent.fix(result.failure);
  result = await env.retest();
}

Let your agent test the app before you do.

TryCase gives your coding agent a disposable Linux desktop to run your app, test it like a user, and send back proof: screenshots, recordings, and logs.

Install TryCase skills in this agent session.

Run:
npx skills add bencsn/trycase-skills --skill trycase-cli --skill trycase-run-linux-app -g

After installation, check whether the TryCase skills are available. Do not ask me to restart this session just for the skills.

If the skills are not available in this session, continue anyway: use web search to find the TryCase docs at https://www.trycase.dev/docs, learn the current CLI workflow from the docs, then use TryCase from the CLI.

Make sure TryCase is installed/up to date, run npx trycase@latest login if login is needed, then use TryCase to test my app and return screenshots, recordings, logs, and any artifacts.

Ask your coding agent to install the skills, or fall back to the TryCase docs if skills are unavailable.

What it looks like

One prompt in, proof out.

agent proof run

You:

How it works

Three steps from prompt to proof.

Install the skills

One command teaches Claude, Codex, Cursor, or any coding agent how to use TryCase.

Ask for proof

The agent gets a disposable Linux desktop with terminal and browser control to run your app and test it like a user.

Review what comes back

Screenshots, recordings, logs, and artifacts return with the answer. The environment is destroyed afterwards.

Example prompts

Ask the agent to prove the work, not just finish the code.

Click a prompt to copy it, then drop it into your coding agent once the TryCase skills are installed.

Ship code your agent already tested.

Install the skills once, then ask for proof on every task.

Get started Open the app