SC 01FeaturesConnect your AI

The video editor your AI drives.

Tell Claude, ChatGPT, Cursor or Codex what you want, and it edits in Frapea while you watch, then hands you back a real timeline you can change.

  1. Add to your AI. Step-by-step for your app →
  2. Open Frapea and click Connect your AI.
  3. Tell your AI the code. Then ask for the video.
Open Frapea
SC 02Sourcing

Stock footage, sound effects and voiceover, built in.

Connect your Pexels, Freesound and ElevenLabs keys once. Your AI searches millions of clips and sounds, works with your ElevenLabs voices, and places what it picks right on your timeline.

Pexels stock video and photos

“Find a slow aerial of a coastline.” Your AI searches Pexels videos or photos, filters by orientation, and drops its pick on your timeline.

▶ Set it up · 2:00

Freesound sound effects and music

Whooshes, rain, a door, a crowd. Your AI filters Freesound by licence, so commercial projects get sounds you can use, and places each one under the frame it belongs to.

▶ Set it up · 2:11

ElevenLabs voiceover

Your AI writes the script, picks one of your ElevenLabs voices and directs the read, calm or excited, then captions it word by word. No ElevenLabs account? Free voices run right in your browser.

▶ Set it up · 2:38
Open Frapea
SC 03The editor

Everything your AI does, you can do by hand.

Under your AI is a full multitrack editor in Chrome, Edge or Brave. Your footage stays on your computer. Open what your AI made and keep cutting.

Before it says it's done, your AI renders frames from the timeline and checks them. How your AI checks its own work →

The Frapea editor: media pool, viewer and a multitrack timeline with video, captions and audio.

Timeline

Unlimited tracks, razor and ripple, snapping that sticks, nested timelines, built the way you already edit.

Everything in Timeline
  • Unlimited video and audio tracks, resizable, with per-track mute, hide and sync-lock
  • Pointer and razor tools, blade at playhead, edge trimming with headroom
  • Ripple delete and ripple insert
  • Snapping that sticks, with the playhead taking priority
  • Linked audio and video that move together until you tell them not to
  • Mark in and out, nested timelines, multiple timeline tabs
  • Speed retiming, and unlimited undo
  • Six transitions — cross dissolve, dip to black, dip to white, wipe, slide and push
  • Adjustable duration and alignment, with equal-power audio crossfades
  • Screen layouts: side by side, top and bottom, picture in picture, 2×2, main and sidebar, three-up

Colour

Colour wheels, curves, .cube LUTs and scopes, with keyframes on every property.

Everything in Colour
  • Transform, crop, opacity, blend modes
  • On-canvas handles — drag the frame, rotate it, crop it, rather than typing numbers
  • Keyframes on position, scale, rotation, opacity, crop and volume
  • Linear, hold and smooth interpolation
  • Exposure, contrast, saturation, temperature, tint, highlights, shadows, levels, vibrance
  • Colour wheels, hue curves, tone curves and .cube LUTs
  • Blur, sharpen, noise reduction, motion blur, grain, vignette, glow, chroma key
  • Scopes, so the grade is a measurement rather than an opinion

Words and captions

Local Whisper transcription with word timing. Delete a word and the picture goes with it. Captions styled in a click.

Everything in Words and captions
  • Whisper transcription with word-level timestamps, at three quality settings
  • Captions generated from the transcript, styled, with presets and group restyling
  • Text clips with outline, background and alignment, and ten animations
  • Text-based editing: delete a word from the transcript and the picture goes with it
  • Silence removal, tight, balanced or loose
  • SRT and VTT sidecars on export
  • Waveforms, per-clip gain and dB keyframe envelopes
  • Linear and smooth fades, track mute and solo
  • Cross-correlation audio sync for dual-system sound
  • AI denoise with a dry/wet mix
  • Ducking under speech, and loudness measured on export

Find any shot

Describe the shot instead of remembering the filename. Semantic search over every frame you own, and face tracking for every face.

Everything in Find any shot
  • Semantic search over your footage — describe the shot, not the filename
  • SigLIP 2 embeddings, so “the shot where she turns around” finds the shot where she turns around
  • Scene sampling that promotes real cuts rather than sampling on a metronome
  • Face tracking: every face in a clip, each with its own path
  • Merge two sightings of one person; bind a mask, label or blur to a head
  • Multicam angle switching
AlsoIn the editor

Motion graphics

  • Animated titles, end cards and infographics as first-class pool items
  • Written as code — spring physics, interpolation, sequences
  • Editable props: change the headline or the colours without touching the code
  • Authored by your agent, or by you, or both

Export

  • H.264, H.265, VP9 and AV1 — whatever your hardware can encode
  • MP4 and WebM, 720p to 4K, or match the timeline
  • Background rendering with progress; the Frapea watermark is one free switch to turn off
  • SRT and VTT subtitle sidecars

You export from the Export panel, in your browser.

Connect your AI and ask for your next video.

Open Frapea in your browser, click Connect your AI, and tell it the code. No installer, nothing extra to pay.

Open FrapeaSee what people make →