If a run looks stuck

Pending means that scene has not been submitted yet. Only one scene runs at a time. Check these in order:

1. Confirm you clicked Parse Prompts first (or Start, which will parse for you). The stats line should show a prompt count, not 0 Prompts.
2. Gemini must be open at https://gemini.google.com. The popup hint should say Gemini tab detected. If it says Open Gemini, open that site, then Start again.
3. Click into the Gemini chat box once, then Start or Resume. The Send button stays disabled until Gemini sees text in the composer.
4. If you just reloaded the extension on chrome://extensions, refresh the Gemini tab, then Start again. The helper script is missing until that refresh.
5. Watch the first row. It should move to ▶ (Waiting for Gemini / Downloading) before the others leave Pending. If nothing moves for 20 seconds, click Stop, refresh Gemini, then Resume.

If Start is blocked with a Pro message, the queue is over 3 prompts on the free plan. Buy Pro or paste a license. That is not a hung batch.

The batch likely paused, stopped, or recovered after a refresh. Completed scenes are never sent again.

1. Read the status line under the progress bar. If it says Paused, Stopped, or “Paused after a reload”, click Resume with Gemini open.
2. If that one scene shows Failed, click Retry on that row, then Resume if the batch is not running.
3. If Gemini refreshed, slept, or Chrome’s Memory Saver dumped the tab, in-flight work returns to Pending on purpose. Refresh Gemini so it is idle, then Resume.
4. If Continue on failure is off, a failed scene pauses the rest. Turn it on in the Batch tab, Retry the failed row, then Resume.

The extension is waiting for a new image, not a timer. Default wait is 180 seconds (Settings → Generation timeout).

1. Look at Gemini itself. If it is still drawing, wait. If it replied with text or a safety message and no image, that scene will fail — Retry with a clearer image prompt.
2. If Gemini is already generating something you started by hand, wait or click Stop in Gemini, then Resume.
3. If it sits there with no spinner, click Pause (or Stop), refresh Gemini, click the composer, then Resume.
4. Slow models: raise Generation timeout in Settings (for example 300 seconds) and Resume.

The prompt ran, but Gemini did not produce a new image the extension could see in time.

• Ask for an image explicitly in that prompt. A chat-only reply will fail the scene.
• Click Retry on that row. Failed scenes do not consume the filename number, so Retry still writes scene_003 if that was scene 3.
• Raise the generation timeout if Gemini is just slow.
• If Gemini’s layout changed and nothing ever detects, refresh the tab. If it still fails, open Help from the popup after an extension update.

Separate scenes with a divider. The extension does not rewrite your text.

Use one of:

• ===PROMPT=== (recommended)
• ---PROMPT--- or ### PROMPT ### or ===SCENE===
• Numbered blocks like SCENE 1: then SCENE 2:

A single pasted paragraph with no divider is treated as one prompt, not many.

Only one batch runs at a time. Use Pause or Stop, then Start or Resume. Do not click Start twice to “speed it up.”

No. The loop runs in the background plus the Gemini tab. Closing the popup is fine. Reopen it to watch progress.

Do not close the Gemini tab, do not let the PC sleep, and do not let Memory Saver freeze that tab.

The current scene goes back to Pending and the batch pauses. That prevents duplicate submits of scenes that already saved.

Wait until Gemini is idle, then click Resume. Completed rows stay completed.

• Pause — finishes the current image, then waits. Resume continues the rest.
• Stop — same for the current scene, then marks the batch stopped. Queue is kept. Resume later.
• Skip — marks that pending scene skipped. That number is left unused so later scene numbers stay aligned with your script.
• Retry — puts a failed or skipped scene back to Pending. Resume if the batch is not running.
• Clear Batch — deletes the queue. Refused while a run is in progress. Stop first.

1. Open https://gemini.google.com in a normal tab (not a screenshot, not another Google site).
2. Refresh that tab after installing or reloading the extension.
3. Click Start Batch again. The extension binds to that Gemini tab.

If you have several Gemini tabs, Start uses the one that is active (or the first Gemini tab it finds). Keep one story chat in front.

Click the Gemini composer once so it is focused, then Resume. If Gemini still looks busy, wait or press Stop in Gemini first.

If you changed the “Conversation” setting to start a new chat and the New chat button was not found, switch it back to Use current chat in Settings, refresh, and Resume.

Settings → Conversation → Start a new chat first begins a fresh chat before scene 1. Or start a new Gemini chat yourself, then Start Batch with “Use current chat.”

Optional Prepend global instruction is off by default. Turn it on only if you want the same character/style line added in front of every scene.

Turn off “Ask where to save each file before downloading”:

• Chrome: chrome://settings/downloads
• Edge: edge://settings/downloads

The extension cannot dismiss that system prompt. Confirm once if it already appeared, then Resume.

They go to your browser Downloads folder:

Downloads/Gemini/<Project Name>/scene_001.png

Project name and prefix are on the Auto Save tab. Change them before Start if you want a new story folder. “Last saved” in the popup shows the newest filename.

That is intentional. Scene 3 does not take scene_003 until it actually saves. Retry scene 3 to write scene_003. Skip leaves that number unused so scene 4 can still be scene_004 if you want the script and files to match.

Auto Save ignores images that were already on the page when the tab loaded. If an old chat dumped many images at once, they are skipped as a bulk load.

The same image re-drawing in the page is usually ignored as a duplicate. A true new render from Gemini is saved as the next number. Use Download Last Image if you only want another copy (_copy suffix) without advancing the counter.

The browser rejected the file. Check Downloads permissions, disk space, and that the Gemini tab is still open. Retry the scene. The next number is only used after a successful save.

Auto Save is free. Batch over 3 prompts per run needs Pro. You can still parse a long script.

1. Click Buy Pro — one-time (or buy on this site).
2. Pay on the checkout page. Check the email you used at checkout for subject Your Gemini Auto Image Saver Pro license.
3. Popup → Settings → paste the GAS1.... key → Activate license.

The payment receipt is not the license. Only the GAS1 key unlocks Batch.

• Search spam for Your Gemini Auto Image Saver Pro license.
• The key is sent to the checkout email, not necessarily the Gmail on Gemini.
• If you used a test From address during setup, only the store owner’s Resend account can receive mail. Real buyers need a verified sending domain.
• Open the thanks page steps, then contact support from the Chrome Web Store listing with your order email if it still never arrives.

Paste the whole value starting with GAS1. — two dots, no extra spaces. A Polar or Lemon Squeezy dashboard key (XXXX-XXXX) will fail; turn those off on the product and use only the emailed GAS1 key.

The license is stored in this browser profile only. Activate again in Edge if you switched browsers.

1. Popup → Auto Save Images ON. Status should say ACTIVE.
2. Refresh Gemini after installing the extension.
3. Generate a new image. Images already in the chat are not downloaded.
4. Use Scan Current Page to see how many generated images are visible. Scan does not download old ones.

The extension tries Gemini’s own download URL, then a full-size hosted image, then a fallback. Compare the file with Gemini’s Download button. If it is still small, generate once more or use Download Last Image after the full render is visible.

Batch pauses when the Gemini tab is frozen or the machine sleeps. Wake the PC, make sure Gemini is a real tab (not discarded), refresh if the page reloaded, then Resume.

In Chrome, pin the Gemini tab or exclude it from Memory Saver. Other tabs (CapCut, YouTube) are fine.

Refresh Gemini once after Load unpacked or after every Reload on chrome://extensions. Edge is the same at edge://extensions.

By design it only runs on https://gemini.google.com/*. It does not read Google passwords or other tabs.