← System records

Media systems · Apple Silicon · Approval workflows

Production Studio

An Apple Silicon app for producing short illustrated explainer videos, where the script, voice, visual plan, and final cut each need a separate human approval.

State
v0.8 release candidate in engineering
Form
Case study
Updated
2026-09-24
  1. Script + evidence
  2. Voice + visual plan
  3. Review gates
  4. Authorized export
01

Problem

Automated media pipelines can obscure provenance, quality gates, and who authorized the final artifact.

02

Actual state

Version 0.8 is a release candidate. Its isolated Engineering Lab runs script and evidence editing, the first approval gate, voice sampling, and narration bound to the chosen sample, and a four-act illustrated study was accepted as a creative benchmark on 24 September 2026. Timing and voice acceptance remain open, and the app has not been configured or adopted for Christian's own production work.

03

System architecture

One production service is the only state-changing boundary; the app, MCP server, jobs, and renderers all call it. Accepted narration sets the timeline, captions come from measured word timing, and each gate is a fingerprinted human decision.

04

Christian's role

Designed the Apple Silicon architecture, production authority model, bounded proposal interface, and deterministic media contracts; defined the acceptance suite and reviewed the implementation against it.

05

Verified evidence

clean-installClean Apple Silicon installation accepted
fixtureSynthetic workflow and rendering acceptance
qualityDecoded-media quality and parity checks
benchmarkFour-act illustrated study accepted as a creative benchmark
parityDeterministic capture: every frame receipted, VMAF 97.7 against the accepted study
06

Safety, privacy, and authority boundary

  • Not configured for Christian's own operations
  • Publishing is unavailable
  • Components are not yet qualified for production
  • No customers, outcomes, or public source release are claimed
07

Current limitation

  • Timing and voice acceptance, component qualification, and the first two short explainers remain open.
08

Next milestone

Deliver two unrelated 30–60 second explainers built from three reusable illustrated mechanisms.