Frami
Frami, the built-in AI assistant
Ask in plain words. Frami plans, uses the same 133 editor tools you drive by hand, and checks the frame it rendered — every change a normal, undoable edit in your History.
On this page
Open Frami
Press Ctrl+J to open Frami. It runs on your plan's cloud models with no API key to paste — or, on a Pro plan, you can point it at your own Claude Code or Codex as its brain (see connect your own agent for driving the editor from outside).
How to ask
Type (or talk) in plain language. Frami keeps a chat per project, so you can start new chats, switch between them, and pick up where you left off after a restart.
- Queue & stop: send more while it works and the messages wait their turn; Stop ends the reply at once, running exports and cut-outs included.
- Attach files: paste, drop or pick pictures, PDFs, Word documents and text files.
- Video links: paste a YouTube or Vimeo link (or a video file) and it watches it — frames across the whole video, the transcript from its captions, and the chapters. Nothing is imported.
- The web: it can search, read pages, and look at a page as a screenshot.
What it can do
Frami can see and hear your edit, then make real edits. A few things to ask for:
| You ask | It… |
|---|---|
| “Remove the long pauses and add captions” | transcribes on your GPU, cuts every silence across the timeline (captions and B-roll stay in sync), lays timed captions |
| “Make the intro punchier” | watches the edit, adds a 3D transition and a glow in one undo step, then renders the frame to check it |
| “Is the music too loud under my voice?” | measures loudness per track in LUFS and ducks the music under speech |
| “Make a vertical version for Reels” | reframes the canvas and every layer to 9:16 |
| “Find where I talk about the mountains” | searches inside the transcript and jumps there |
| “Put the title behind me” | cuts you out of the shot on your computer and slips the title between you and the background |
It can render a frame exactly as the export will look, read a whole edit as one contact-sheet image, look inside footage, read the transcript, and measure loudness, peaks and silences — so it verifies its own work.
Approvals & undo
An agent with hands only helps if you stay in control. Itsframe keeps every edit visible and reversible.
- Every AI edit is a normal, typed command in your History, marked as made by the AI.
- Undo all reverts a whole answer;
batch_editgroups several changes into one all-or-nothing undo step. - Nothing happens out of sight: the chat shows the model's thinking and each step as it is written, run and finished.
- Steps that run code in Blender wait for your explicit approval before they run, and Frami asks before it looks at your screen.
Generation features
You can generate straight into your media bin:
- 3D & motion graphics: extruded 3D titles and looping backgrounds in Blender, and HyperFrames HTML + GSAP motion graphics.
- Images & video: with your own OpenAI key (GPT Image) or Google key (Gemini image models, Veo video) — cost is shown and confirmed before a video runs.
- Voice, SFX & music: voiceovers with ElevenLabs or Gemini TTS (transcribed for captions), plus ElevenLabs sound effects and music.
- Anything from a connected MCP service (Higgsfield, Runway, Replicate, fal.ai or your own) lands in your media bin.
Credits
Only generation that Itsframe runs for you — through Frami on your plan's cloud models — spends plan credits. To use it, sign in from Integrations › Account & credits to link your device.