AI Content Transformation Platform
Helps content teams turn long transcripts into post drafts for review before anything is scheduled.
Project overview · 33 seconds
I built a transcript-to-draft publishing product
I built a transcript-to-draft publishing product that turns long-form audio into post drafts for review.
Content teams can record hours of valuable material, then lose momentum because useful ideas stay buried in long transcripts. Manual repurposing is slow, voice is hard to maintain across posts, and finding the most shareable moments requires judgment.
I turned transcripts into drafts people could review
Vox Prismatic turns long-form transcripts into draft posts, then keeps a person in control of review and scheduling before publishing.
The product needed to extract useful moments, preserve the speaker’s voice, and connect generation to the rest of the publishing process.
The workflow separates extraction, drafting, and approval
Turn source transcripts into reviewable drafts.
The workflow treats transcripts as a source that needs extraction, shaping, and review before publishing.
Show progress while longer jobs run.
Progress updates show where a transcript is in the process, while review and scheduling controls carry the draft beyond a single AI response.
The interface keeps review before scheduling
The interface shows the progression: a long transcript becomes candidate ideas and drafts, while scheduling remains a separate human decision.
I connected transcript analysis to the publishing workflow
Break long transcripts into workable sections.
The system chunks long transcripts, extracts candidate insights, and turns them into drafts someone can review instead of mining the transcript by hand.
Carry approved drafts into scheduling.
Opening frameworks, scheduling controls, timezone-aware publishing, and progress updates carry the draft from generation into a usable publishing process.
The workflow ends with a reviewable draft
Vox Prismatic turned a manual transcript-to-post workflow into a system for extracting, drafting, reviewing, and scheduling content.
Long inputs became reviewable drafts.
The system chunks transcripts, extracts candidate insights, and turns them into post drafts for human review.
Publishing stayed human-led.
The workflow kept review and approval in the loop instead of handing publishing to a generic content generator.
Scheduling became part of the system.
Opening frameworks, scheduling controls, and timezone handling made the product an operating tool.
Long-running AI work had visible progress.
Real-time updates made processing feel less like a black box while the AI workflow ran.
Have a product priority that needs an owner?
I join B2B SaaS teams on contract to ship the features, internal tools, and integrations already on your roadmap.