How sidegig.directory is made

What is in the catalog

The catalog currently holds 65 published side hustles, 2327 accepted facts and 8142 supporting quotes and data points.

How claims are gathered

We collect public videos and their transcripts from creators who talk about side hustles, plus public datasets. Each piece of content is read by a language model that proposes structured facts (for example typical pay, start-up cost or time commitment). Every proposed fact is stored together with the exact quote or data point it came from, and the source link.

How claims are scored

A fact is published only when it has supporting evidence. Confidence rises when independent sources agree and falls when they conflict. Facts with too little evidence stay off the page. Each fact shows its confidence, the date it applies to, and numbered links to its sources.

How pages are written

Tables, figures and quotes are taken directly from the stored facts. Short explanatory prose is written by an AI model from those facts only, and automated tests check that every sentence is supported by a cited fact before a page is published. Pages are generated on the server; nothing needs JavaScript to read.

AI disclosure

AI models are used to extract facts from sources and to write prose on these pages. They are not used to invent figures. Treat what you read as a well-sourced starting point and check the linked sources before you rely on it for a financial decision.

Public datasets

• BLS Employment Projections, current release 2025-35
• BLS Occupational Employment and Wage Statistics, current release May 2025
• CareerOneStop License Finder, current release 2024-10-14
• Census Nonemployer Statistics, current release 2024
• O*NET 31.0 Database, current release 31.0

Use it in Claude

The catalog is also available as a connector: see how to add it.

Corrections

If a fact is wrong or out of date, tell us and we will review it. Our terms explain how the content may be used.