ProjectCase Study

ThreadScribe Studio

Trustworthy AI transcript interaction, built local-first

A local-first transcript workspace that keeps raw speech, AI drafts, provider state, retention, and failure recovery visible throughout the workflow.

Year2026

ExperienceGuided product walkthrough

CategoryTrustworthy AI + frontend systems

Next.jsTypeScriptSQLiteAI evaluation

ContextThe work

Keep the source close to the draft.

ThreadScribe treats raw speech, provider state, transformed output, retention, and recovery as first-class parts of the same product. The useful artifact is only trustworthy when the path that produced it remains inspectable.

The problem

Turn private speech into useful written artifacts without hiding where processing happens, replacing source material silently, or presenting sample output as live-model intelligence.

The user

Solo builders who record implementation notes, bug reports, planning thoughts, and short voice memos that need to become durable, searchable work.

My role

Independent product direction, UX/UI, Next.js implementation, local persistence, AI evaluation, privacy review, accessibility, and release evidence

ProofDecisions

The choices that made the product hold together.

Kept timestamped raw segments

Kept timestamped raw segments beside transformed drafts so the source remains inspectable.

Made sample, local, and

Made sample, local, and cloud provider modes explicit, with external calls disabled by default.

Stored sessions, edits, outputs,

Stored sessions, edits, outputs, tags, and exports in a local SQLite library with verified deletion behavior.

Preserved audio previews and

Preserved audio previews and prior work across cancellation, worker failure, malformed output, and retry.

ScreensSelected work

The interface, in context.

ThreadScribe raw timestamped transcript view — Timestamped source segments stay available beside the transformed work.

ThreadScribe transformed transcript output — The output is useful without pretending a sample transform is a live model result.

ThreadScribe recoverable worker failure state — Failure preserves prior work and gives the user a clear recovery path.

ReleaseProof and boundaries

What the build makes visible.

Verification

Five labeled sample cases pass 45 of 45 deterministic checks across cleanup, notes, actions, empty input, and instruction-like transcript text.
Thirty-seven application tests pass alongside privacy samples, keyboard workflow checks, production build, and automated accessibility scans.
One permitted 8.63-second clip completed with local faster-whisper in 2.068 seconds warm on the recorded demo machine.

Limitations

The public walkthrough uses deterministic sample transforms. It is not evidence of general model quality.
The local Whisper result covers one short English clip on one machine and does not establish production-scale accuracy or latency.
Speaker labels and segment confidence remain exploratory; hosted sync, sharing, and compliance certification are outside the project scope.

MoreCurrent case studies

Explore the other current projects where product thinking is carried through interface, architecture, testing, and release.