Public roadmap
What we're building next.
Ordered by when it ships, not by wishful thinking. Every item ties to a real user need. Timelines are targets; we ship when it's good, not when it's due.
◆
Delivery-scored transcripts
Per-word energy, pace, and pitch, z-scored against each speaker's own baseline in that video.
Live
⭐
Best-Take Autopilot
Detects repeated takes, ranks them on delivery, flags abandoned attempts, always defers to you on ties.
Live
✂️
Clip mining with overlap ranking
Long-form footage becomes short-form candidates, ranked by where hook and delivery both land.
Live
👁
Facial engagement + body language scoring
Vision analysis on flagged moments. Observable signals only, never inferred internal states.
Beta
📝
Emphasis-aware styled captions
Captions animate on the words the speaker actually hit. Slow emphasis renders in its own style. The visible face of the engine.
Building
📤
Auto-publish + scheduler
Direct posting to TikTok, Instagram Reels, YouTube Shorts, Meta, X, and LinkedIn with per-platform captions and a content calendar.
Building
🎨
Brand kits and templates
Custom caption styles, fonts, colors, logo placement, intro and outro cards. Multi-brand support for agencies.
Building
🔲
Auto-reframe engine
16:9 to 9:16 with speaker tracking. Keeps faces centered even as they move through the frame.
Planned
✨
AI titles, hooks, and descriptions
Generated per clip, informed by your delivery data. "Open on the energy spike at 00:37."
Planned
🎙
Audio enhancement suite
Studio-sound cleanup, loudness normalization, de-noise, per-speaker leveling.
Later
📊
Performance analytics + clip feedback loop
Track published clip performance and correlate with delivery signals. "Your slow-emphasis clips outperform energy-spike clips 2:1." Nobody else can offer this because nobody else has the signal data.
Later
🗣
Speaker delivery report
Per-episode coaching. Filler-word trends over time, pacing analysis, energy map, "your strongest 10 seconds." A second product line hiding inside the engine.
Later
🌐
Translation + AI dubbing
Auto-translated captions first. Voice-cloned dubbing later with explicit consent controls.
Later
🎞
AI b-roll and visual enrichment
Suggested stock and generated b-roll placed where delivery sags, not where keywords match. Credit-metered on paid tiers.
Later
🎤
AI voice regeneration + lip-sync
Re-record a misspoken word with your own voice model and regenerated mouth movement. Deep consent controls; opt-in per project.
Later
🏢
Agency white-label + client workspaces
Deliver under your own brand. Per-seat + per-client pricing. Enterprise-grade access controls.
Later
🔌
Public API and MCP server
The delivery-analysis engine as an API. Model Context Protocol server for Claude Desktop, Cursor, Zed, and other agent environments.
Later
How we prioritize. Anything on this roadmap has to serve one of the three product jobs: delivery-aware editing, the full trim, or the volume of short-form output. If a feature doesn't ladder up to one of those, it doesn't ship. We reprioritize based on eval-set findings and user feedback, not competitor pressure. If something you need is missing, tell us at
help.