Coming soon Revela
Raw footage to publish-ready reels: planned by AI, rendered on your machine
Application
Revela
Raw footage to publish-ready reels: planned by AI, rendered on your machine
This production is a private summary: description, learnings and stack are shown, the code stays private.
Draft. Written from the program notes and pending the author’s review.
- Status
- Coming soon
- Type
- Application
- Running since
- January 2025
- Role
- Architect and sole engineer
- Domains
- video editing, creator tools
- Cast
- Python, FFmpeg, Litellm
Context and problem
Short-video planning and editing: karaoke captions, cuts, reframing and music matched by BPM.
Assumption: creators and small brands spend most of their editing time choosing and ordering clips rather than rendering, and the tools that automate the choice upload the footage to someone else.
Revela turns raw footage into publish-ready reels: a model plans the cut through a gateway, and the render runs on the creator's own machine, so the footage stays local.
Screens
How it was built
Built in 6.8 h · USD 170.93 of AI · 488 tests · 4 audits
Source: Trysac/10-revela docs/development-report.md (lines 11, 14, 31), dated 21 November 2026.
What the project taught
- Separating the plan from the render lets a cloud model decide the cut while the footage never leaves the machine.
- Routing every model call through a gateway keeps the planner independent of a single provider and makes cost visible per run.
- Rendering on a CPU-only laptop sets the real budget: the planner must produce a cut that renders in acceptable time, not the best possible cut.
Case study in preparation
The full case study adds these sections when it is published:
- Constraints
- Architecture diagrams (C4 context and containers)
- Three key decisions
- What went wrong
- Results with dated evidence
- Demo video
- What I would do differently
In the project mesh
Highlights
local rendering, LLM gateway, planning