Solo Founder Playbook draws on 101 Starter Story founder interviews. Here is the actual workflow used to turn long videos into a small, inspectable knowledge base.
1. Collect the videos
The pipeline uses the YouTube Data API to list videos from a channel and excludes clips shorter than five minutes. Each video gets a record before any AI analysis begins.
2. Download transcripts
Transcripts are fetched through youtube-transcript.io. The step is incremental: on another run, already downloaded videos are skipped.
3. Analyze each interview
Claude reads the transcript and organizes it into 11 fixed topics, including idea generation, validation, launch, marketing, growth, monetization, and failures. A fixed structure makes later comparisons possible.
4. Refine the notes
A second pass produces a short product description and turns long passages into concise, usable summaries. The aim is to preserve the meaning of each interview while making it easier to compare.
5. Compare across interviews
Claude looks across the analyzed interviews for recurring strategies and tools. The resulting pattern files and human-readable insights supply the knowledge base for the six skills.
What the output can and cannot tell you
The published set contains more than 80 recurring patterns across 11 categories. Frequency shows what appeared in these 101 interviews. It does not show whether a tactic caused a founder's outcome, and this YouTube interview sample is not representative of every startup.
The pipeline code and README explain how to run each step and adapt the categories for another channel. The README estimates roughly $5–10 for transcript retrieval and $15–30 for Claude API processing per 100 videos; actual cost depends on video length, model pricing, and service plans.
The end product is Solo Founder Playbook. For a concrete example of the patterns, see the failure-mode analysis.