Open-source skill turns podcast transcripts into publishable articles, cutting 35% with zero detail loss
oran_ge · x · 2026-10-07
The author open-sourced an MIT-licensed Agent Skill, podcast-to-article, built after struggling to turn their podcast Next Token into readable articles: raw transcripts read like meeting minutes, while AI polish strips out numbers and details.
Core idea: re-axe the content. Transcripts are organized by speaker and time; articles are organized by topic and judgment. The skill rebuilds the structure by topic while enforcing a two-way rule: dismantle the speaker scaffolding, but never cut numbers or direct quotes—only drop off-topic tangents, news name-drops, and self-promotion.
Measured results: a 5,730-character transcript became a 3,703-character publishable piece (35% reduction, 57→47 paragraphs), and style-check scripts showed zero AI-isms—no dashes, contrastive sentence pairs, summary tone, or ordinal headings.
Install by cloning the repo and symlinking it into your agent's skills directory, pointing at SKILL.md.
More from coding & agent
- Herald OS goes open source: an agent-native OS where the AI is the interface, not an app — gekobraa · 2026-10-07
- MCP observability tool adds per-tool health grades to catch silent agent failures — Thirumalaiboobathi · 2026-10-07
- Hermes Gadget SDK sparks community builds on watches and old phones three days after release — Teknium · 2026-10-07
- Glasser aggregates 1,558 data APIs from 29 providers for pay-as-you-go AI agents — goyalshaliniuk · 2026-10-07
- MIT paper: a minimalist agent loop that passes history as code variables beats Letta and ACE at half the cost — rohanpaul_ai · 2026-10-07
- A 3Blue1Brown-Style Explainer Rendered End-to-End in Rust by franken_manim — doodlestein · 2026-10-07