Home / Terms
Terms of use
Short version: it’s free, it’s information and not advice, our own data is yours to reuse with credit, and other people’s data stays theirs.
What this is
- Glitch in the Credit is a free website and daily email about stress in the debt that pays for AI data centers. One person runs it.
- Nothing here is investment, legal or tax advice, or a recommendation to buy or sell anything. It isn’t a timing tool.
- The readings are judgment calls. An AI model grades each signal every morning from public news, filings and data. They can be wrong, late or incomplete, and some rest on secondhand reports. Every reading links its source so you can check it.
- We work to keep the numbers right and we publish our corrections, but we don’t promise that the site, the data, the API or the email is accurate, complete or always available. You use them at your own risk.
- As far as the law allows, we aren’t liable for any loss that comes from using or relying on them.
Reusing our data
- Our own work is free to reuse under CC BY 4.0: the stress index, the layer and signal scores and statuses, the daily archive, and our GPU cost inputs and floors. Credit “Glitch in the Credit” and link to the page you used.
- The government data we pass on (SEC, Federal Reserve, OFR, FFIEC) is public. Credit the agency it came from.
- Please don’t present our numbers as your own, or suggest that we endorse you or your product.
Other people’s data
- Some numbers on the pages belong to others: stock and fund prices, GPU rental index prices, the VIX. We show them with credit so the picture is complete. They aren’t ours to license, so they aren’t in any download, and we ask that you don’t copy them from our pages in bulk. Get them from the source, listed under where the data comes from.
- Company names and logos belong to their owners. We use them only to say who is who. We aren’t affiliated with or endorsed by the companies we write about.
The API and automated readers
- The files under /data/ and api/latest.json are open. No key, no sign-up.
- The full dataset needs a free key, which comes with the daily email. Each key is for one subscriber: 10 requests a minute and 500 a day. Please don’t share or publish it.
- We may change the limits, and we may switch off a key that goes far past them or is used to get around these terms. A key stops working within a week of unsubscribing.
- Crawlers and AI agents are welcome. robots.txt and llms.txt say what to read and what to leave.
The newsletter
- It’s free: one email each morning. Every email has a one-click unsubscribe.
- What we keep about subscribers, and how to have it deleted, is on the privacy page.
Mistakes and changes
- If you find a mistake, email [email protected]. Corrections are logged on the data status page.
- If these terms change, the new version goes on this page with a new date.
Last updated