New media tools in MCP, Grok Imagine 1.5, and improvements to the Story Engine

Hi there,

In the last email, we made sure you have the ability to use your claude or chatgpt sub to write or edit a whole film into Frameliq. Now it can generate the media too: Claude or ChatGPT (AI assistants) can generate your storyboard stills and your Composer video shots from the same chat.

It tells you the cost before anything runs, and nothing starts until you say yes.

Also in this update: shots are now planned as a sequence, so each one picks up where the last one ended. That is section 5 below.

1. Stills and video, generated from the chat

Ask for the stills for a scene, or the video for a set of shots. The assistant quotes the cost in Media credits and the balance you would have left, waits for your yes, starts the generation, then saves each result onto its storyboard panel or Composer shot. Stills take under a minute and video takes a few minutes. Any Frameliq tab you have open updates as the results land.

Try it with two messages:

• "How many Media credits would it cost to generate the stills for scene 1? Tell me the cost and my balance before you start."
• "I approve that cost. Generate the stills for scene 1, then show me the results."

2. It sets up the whole shot, not just the prompt

Before, your AI assistant wrote a motion prompt and you did the rest in Composer. Now it can choose the video model, mode, resolution and length for each shot, attach reference images by frame, character or place, and give a speaking character their voice. You can still open any shot in Composer and change it before it renders.

3. Spending stays under your control

• The assistant can spend at most 300 Media credits per request and 1,500 in any 24 hours.
• A quote lasts 15 minutes, works once, and is void if you edit the frame or shot afterwards.
• A generation the provider rejects is not charged.
• It is paid the same way as generating in the app: your included Media credits, or your own fal key. One Media credit is $0.01 of generation.
• Ask "how many credits do I have left?" at any time.

Character and place portraits, voices, music, lip sync and the final export are still done inside Frameliq.

4. Grok Imagine 1.5 is in Composer

xAI's two new video models are in the model picker: Grok Imagine 1.5 and Grok Imagine 1.5 Lite. Both make picture and sound together, in takes from 1 to 15 seconds at up to 1080p. The full model also has a References mode that takes up to 7 images at 480p or 720p. Lite costs less per second, so it suits quick drafts. Both run on included Media credits.

5. Improved Story Engine: Shots that follow each other (better continuity)

Until now each shot was planned and generated on its own, which is where jump cuts, repeated action and drifting characters came from. A storyboard now carries a scene plan that the stills, Composer and your assistant all read:

• Staging. Each scene records who is on screen and where they stand, and each shot says how it joins the one before it: a continuous move, a reverse angle, a cut-in or a jump in time.
• Shot length that fits. The length of a shot is worked out from its lines and its action, and a take the model made too long is trimmed back to plan.
• Frames that carry over. A shot that continues the previous one opens on the frame that take ended on.
• Frame check. Frameliq can compare a still against the cast and the scene before you spend video on it. It uses Story credits, only gives advice, and never starts a generation.
• Dialogue fit. In Composer's Audio tab, a shot shorter than its recorded lines is flagged, with "Extend the shot" or "Split the lines" to fix it.
• From your assistant too. Claude and ChatGPT plan with the same staging rules, and can pull up each still beside the cast and the plan to check it before any video is generated.

Storyboards you planned earlier stay as they are until you plan them again.

Smaller things you'll notice

• ChatGPT setup is shorter. Add Frameliq from Plugins → Add → Create MCP App. Most accounts no longer need Developer mode. If you have already added, you need to just refresh tools from your Frameliq plugin settings in ChatGPT / Claude, else the new tools will not show up or work.
• When you import a storyboard into Composer, you choose the video model for the new shots in a dropdown.
• Every Import storyboard button now opens the frame picker, so you choose which frames come across.
• New guides on the blog: Can ChatGPT make videos, Can Claude make videos, and a filmmaker's guide to Seedance 2 and 2.5.

No fal key? You don't need one

Most people who stall in Frameliq stall at the same step: images and video ask for a fal key, and they don't have one. There are two ways past it:

• Creator Plus, $29/mo: 800 Media credits a month are included, so stills, video and voices work with no fal account, in the app and from your assistant.
• Creator, $19/mo: add your own fal key in Settings and pay fal directly for what you generate.

Both include 500 Story credits a month and the Claude and ChatGPT connection.

Hit reply and tell us what your assistant made. We read every reply.

Happy filmmaking,

The Frameliq Team