jackwener/OpenCLI · error · CommandExecutionError

Midjourney did not expose a Describe action for the uploaded

Error message

Midjourney did not expose a Describe action for the uploaded image

What it means

This CommandExecutionError is thrown when the browser automation could not find a visible, clickable 'Describe' button or menu item on the Midjourney web UI after uploading an image. The library clicks 'Describe' via page.evaluate by scanning all button/[role=menuitem] nodes for exact text 'Describe' with nonzero width; if none is found, the Describe workflow cannot proceed and this error aborts the command.

Source

Thrown at clis/midjourney/describe.js:89

      const button = card?.querySelector('button');
      if (!button) return false;
      button.setAttribute('data-opencli-describe-menu', '1');
      return true;
    }, sourceUrl);
    if (!menuMarked) {
      throw new CommandExecutionError('Could not open the uploaded image action menu for Describe');
    }
    await page.click('[data-opencli-describe-menu="1"]');
    await page.wait(0.4);
    const startedAt = new Date().toISOString();
    const clicked = await page.evaluate(() => {
      const button = [...document.querySelectorAll('button,[role="menuitem"]')]
        .find((node) => node.textContent?.trim() === 'Describe' && node.getBoundingClientRect().width > 0);
      if (!button) return false;
      button.click();
      return true;
    });
    if (!clicked) throw new CommandExecutionError('Midjourney did not expose a Describe action for the uploaded image');

    const deadline = Date.now() + timeout * 1000;
    let prompts = [];
    while (Date.now() < deadline) {
      const groups = await page.evaluate(() => {
        const visible = (node) => {
          const rect = node.getBoundingClientRect();
          return rect.width > 0 && rect.height > 0;
        };
        const found = [];
        const markers = [...document.querySelectorAll('div')]
          .filter((node) => node.children.length === 0 && node.textContent?.trim() === 'Describe' && visible(node));
        for (const marker of markers) {
          let root = marker.parentElement;
          for (let depth = 0; depth < 16 && root; depth += 1, root = root.parentElement) {
            const rows = [...root.querySelectorAll('p')]
              .filter(visible)
              .map((node) => node.textContent?.trim().replace(/\s+/g, ' ') || '')

View on GitHub (pinned to 49907e53dc)

Solutions

1. Re-run the command after confirming the image upload completed and the Describe option is visible in the browser session
2. Re-authenticate/refresh the Midjourney cookie session and retry, since an expired session can render the UI without actions
3. Check whether Midjourney renamed the Describe action after a UI update; update the selector/text match in clis/midjourney/describe.js if needed
4. Verify the uploaded file is a supported image type and size, then retry
5. Increase the wait before searching for the button so the page/menu finishes rendering

Example fix

// before
const button = [...document.querySelectorAll('button,[role="menuitem"]')]
  .find((node) => node.textContent?.trim() === 'Describe' && node.getBoundingClientRect().width > 0);
// after (tolerate partial labels)
const button = [...document.querySelectorAll('button,[role="menuitem"]')]
  .find((node) => /describe/i.test(node.textContent || '') && node.getBoundingClientRect().width > 0);

// Verify the Describe action exists before invoking the command
const hasDescribe = await page.evaluate(() =>
  [...document.querySelectorAll('button,[role="menuitem"]')].some(
    (n) => n.textContent?.trim() === 'Describe' && n.getBoundingClientRect().width > 0,
  ),
);
if (!hasDescribe) throw new Error('Describe action not visible; check upload/session before running.');

function isDescribeAction(node) {
  return (
    node instanceof HTMLElement &&
    node.textContent?.trim() === 'Describe' &&
    node.getBoundingClientRect().width > 0
  );
}

try {
  await run('midjourney', 'describe', imagePath);
} catch (err) {
  if (err.message.includes('did not expose a Describe action')) {
    // re-authenticate, confirm upload rendered, and retry once
  } else throw err;
}

When it happens

Trigger: After uploading an image, document.querySelectorAll('button,[role=menuitem]') contains no node whose trimmed textContent is exactly 'Describe' with getBoundingClientRect().width > 0 — e.g. the upload silently failed, the Describe menu did not render, the UI labels changed, or the element is hidden.

Common situations: Midjourney UI redesign renaming or moving the Describe action; the image upload failing or being rejected (wrong file type/size) so no Describe option appears; the page not fully loaded or not logged in; popup/overlay blocking the menu from being visible; running in a headless viewport where the element is collapsed.

Related errors

• Midjourney Imagine composer was not found.
• Could not switch to ${wantModel} model
• Claude composer is not available on the current page.
• Codex sidebar extraction returned an invalid payload.
• Could not resolve a stable Codex conversation identity.

AI-assisted analysis of jackwener/OpenCLI@49907e53dc (2026-08-29). Data as JSON: /api/errors/f4777cb0d471954d. Report an issue: GitHub.