To update from the panel, open Settings ▸ About ▸ Check for updates and click Download update, or download the file above. Then install it with the ZXP Installer and restart After Effects. Your settings, keys and history stay where they are.
Problem reports
Settings ▸ Options ▸ Save Problem Report makes one file with what the Summon team needs to look into a problem: up to 30 days of the Log, your Summon, After Effects and system versions, your settings and the models you added. Your fal.ai key and license key are never in it.
Every failure message shows a Problem ID, which is also in the Log. Quote it when you write to support@summonfx.com.
The plain-words line at the top of a failure message now comes from the error itself. When fal.ai rejects a run after it has started, the message and the Log show fal.ai's reason as a sentence instead of raw code.
Prices
Until your first run at the settings you pick, such as resolution, sound, size and quality, estimates use fal.ai's listed price for those settings when one is available. Then they follow what fal.ai actually charged you. On Image, Edit Video and Lipsync, a charge very different from your earlier runs at the same settings waits for one more run to confirm it.
"Why this price?" on the Image, Edit Video and Lipsync tabs says where the estimate comes from.
MiniMax H3, Seedance 2.5 and Seedance v2.0 Fast (from video) include your clip in the estimate, because fal.ai bills them for your clip as well as the video they make. The estimate shows just the amount, and "Why this price?" shows the calculation.
Estimates follow fal.ai's rounding, for example ElevenLabs Dubbing, which fal.ai bills per started minute.
Models
New built-in models. Image: GPT Image 2.5 Sunburst and Flare, each with an Edit version, Seedream 5.0 Pro and Seedream 5.0 Pro Edit, and Kling Image O3 Edit. Animate: MiniMax H3 Max, Seedance 2.0, LTX 2.5 Fast, Wan 3.0 Prime and Wan 3.0 Prime Ref→Video, Veo 3.1 Fast First + Last Frame and FLUX 3 First + Last Frame. Edit Video: FLUX 3 Fast Edit Video. Lipsync: Creatify Aurora and OmniHuman 1.5.
Some older models are no longer built in. Any of them can still be added from the Model Manager.
Settings ▸ Options ▸ Model menus sets the order of the models within each group of the menus: Summon's order, name, price, or recently added.
Models you add from fal.ai get more of their own settings, such as duration, sound, quality and large sizes.
Kling O3 Omni (from video) needs a clip that shows a person's face. The line under the model menu says so.
Working in the panel
Add to Timeline and Add to Project wait for a slow download to finish and show the progress, instead of showing an error after a minute.
An upscaled result works like a generated one: it shows in Result and History, and it goes into your project when you choose Add to Project or Add to Timeline.
Clicking X on a running generation asks first, because fal.ai may still charge for it.
The line under the player shows the video's real length.
Activity drawer: a pinned drawer no longer opens by itself when Summon starts, the pin shows clearly when it is on, and the buttons for a selected take all fit.
Lipsync: capturing the playhead again into the photo slot now shows the new frame.