Skip to content

Common Issues

Extension not listed in Window > Extensions

Section titled “Extension not listed in Window > Extensions”
  1. Check the extension folder exists:
    ~/Library/Application Support/Adobe/CEP/extensions/com.fittoframe/
  2. Enable unsigned extensions:
    defaults write ~/Library/Preferences/com.adobe.CSXS.11.plist PlayerDebugMode 1
  3. Restart Premiere Pro completely — quit and relaunch, not just close and reopen
  • The local Node.js server may not have started. Wait a few seconds — the plugin auto-starts the server on panel open
  • Check if Node.js is installed: open Terminal and run node --version. If not found, install Node.js from nodejs.org
  • If the panel stays blank for more than 10 seconds, close and reopen it from Window > Extensions

The local backend server lost connection. Click the Reload button in the error message. This triggers a full recovery sequence: health check → restart if needed → wait for healthy → retry.

Some models experience queue times during peak usage on fal.ai. This is normal for popular models — especially high-quality video models.

  • Wait — queue times vary from seconds to a few minutes
  • Switch models — try a faster model (e.g., LTX 2.3 instead of Kling v3 Pro)
  • Check fal.ai status — if all models are slow, fal.ai may be experiencing high load

If a generation stays queued for an extended period, it automatically moves to background processing.

Heavy models (Kling v3 Pro with audio, Veo 3) can take 5–15 minutes. The progress button updates its message over time:

  • “Generating…” → “Sit back and relax…” → “Still working on it…” → “This model takes a while…”

If the generation exceeds the expected time, it moves to the background automatically. Check the Active Generations Panel for status.

Your fal.ai account has run out of funds. Add credits at fal.ai/dashboard/billing. After topping up, the next Generate click shows a confirmation: “Have you topped up?” — click yes to proceed.

The button stays disabled until all requirements are met. Hover over it for a tooltip explaining what is missing:

  • “Add a prompt to generate” — the prompt field is empty
  • “Select media from your timeline” — the model requires an image or video input
  • “Add your API key” — no fal.ai API key configured
  • Validation error on the media card — the selected media does not meet the model’s requirements

”Image too small” / “Image too large”

Section titled “”Image too small” / “Image too large””

The selected image does not meet the model’s dimension requirements. The error message shows the exact constraint: “Image too small (640x480, min 1024x768).”

Use a larger or smaller image, or try a different model with less restrictive requirements.

The selected file exceeds the model’s file size limit. Try:

  • Reducing the image resolution before selecting it
  • Using a more compressed format
  • Selecting a different file

Premiere Pro sometimes returns 0x0 for certain image formats. modelBridge reads dimensions directly from the file as a fallback. If the card still shows 0x0, try re-importing the image into Premiere Pro.

The model requires an input image or video. Select a clip on the timeline, select an item in the Project Bin, or drag a file from Finder.

Some fal.ai models are gated — they require approval from the model creator. The error dialog explains the restriction type and how to request access. Access denials are cached for 24 hours; after getting approval, save a new API key or wait for the cache to expire.

The model’s schema may have been cached with outdated information. Model schemas refresh automatically (5-minute TTL), but you can force a refresh by removing and re-adding the model.

fal.ai does not publish pricing for this model. You can still generate — check your fal.ai dashboard afterward for the actual charge. See Costs & Pricing for details on pricing confidence levels.

Estimates are based on the best available pricing data but may not capture all parameter-dependent surcharges. The Billing tab shows both estimated and actual costs for comparison.