How it works
How Log reads a panic log
Log compares the panic string, the model, and the sensor codes with a table of published repair notes. You get a starting point and a confidence level.
What you give it
Paste the panic text, or upload a .ips, .txt, .log, or .json file. The reader accepts one file up to 1.5 MB. That size limit is separate from the pricing draft.
A shop account is optional for reading the guide pages. The upload form on the home page asks you to sign in, and each shop sees only the logs it uploaded.
What the reader looks for
The parser takes the panic string, the product identifier, missing sensor names, a watchdog process, and SMC sensor-array codes. Rules in the rule table are ranked.
The result shows the part, the confidence, the next check the rule suggests, and a link to the public page that rule cites.
Warnings stay on the result
Unknown code bits, a cut-off read, and a code with no row for that model lower the confidence. An oversized SMC code is kept as text and marked unknown, so Log does not round it into a nearby part.
Those warnings are part of the result for every shop. The pricing draft does not sell a quieter warning or a different rule.
What is still a slot
The schematics screen is reserved and empty. Log does not open a work order. Saved histories, batch reads, branded exports, and team tools are described on the pricing page as planned work.