How ExplainMyCar works
Where our information comes from, how we decide what is good enough to publish, and what we deliberately refuse to do.
The short version
We hold 9,533 generic OBD-II trouble codes. 122 of them carry full diagnostic guidance that a person wrote and a second reviewer checked. The rest are searchable and carry their standard definition and structured relationships, and they are explicitly marked noindex so we are not asking search engines to rank pages we have not finished.
That gap is deliberate and we would rather show it than hide it.
Where the code definitions come from
The generic code set is defined by two standards, SAE J2012 andISO 15031-6. Those documents are copyrighted, so we do not reproduce their text. We use the OBDexopen database, whose data is dedicated to the public domain under CC0-1.0, for code identifiers, titles, and the structural relationships between codes, components and symptoms.
Every imported record is stored with a pointer to the exact archived snapshot it came from, including a content hash. If you ask us where a fact came from, we can answer with a specific file rather than a shrug.
Where the vehicle data comes from
VIN decoding uses NHTSA's vPIC database. Recall information comes fromNHTSA's recalls API. Both are US government data. We query them live rather than caching a stale copy, and when they are down we say so rather than showing you nothing.
How the diagnostic guidance is written
The plain-English explanations, the check-this-first ordering, the diagnostic procedures and the cost ranges are written for this site. They are not scraped, and they are not a reworded version of somebody else's article.
Every code that we publish guidance for goes through a second pass whose only job is to find errors in the first: inverted circuit-test logic, a code described as the wrong fault, a repair recommended without a confirming test, severity that is too high or — much more dangerous — too low, and driveability guidance that makes a promise it cannot keep. That review pass is adversarial by design. On our first run it found fifteen errors serious enough to block publication, including a code that had been documented as the wrong fault entirely.
What we will not claim
We do not claim credentials we do not hold. There is no ASE-certified review board here, and you will not find a stock photo of a mechanic with a made-up name attached to an article. If that changes, we will say so and name the person.
We also do not:
- Publish invented probabilities. You will never see "P0420 is caused by a bad converter 73% of the time" here, because nobody knows that number.
- State vehicle-specific facts — torque values, connector pinouts, component locations — that vary by manufacturer, unless we have a source for that specific vehicle.
- Reproduce copyrighted repair procedures, OEM manual text, or competitor articles.
- Generate a page for every possible code-and-vehicle combination just because the combination exists.
How we decide what to publish
A code becomes eligible for search indexing only when reviewed content exists for it and that content scores above our quality threshold. The score is not word count — padding a page does not make it better. It measures whether the things a reader actually needs are present: a plain answer, an explanation of how the fault was detected, an ordered list of what to check first, a real procedure, repairs that are conditional on a test, and the mistakes people commonly make.
If the person who wrote a code's content told us they were unsure about it, that code stays out of the index regardless of how complete the page looks. Honest uncertainty beats a confident page that is wrong.
The distinction everything here rests on
A trouble code is a detection, not a diagnosis.
A code tells you what a control module measured and that the measurement fell outside the range it expected. It does not tell you which part failed. P0420 does not mean your catalytic converter is bad — it means the computer measured catalyst efficiency below its threshold, and a worn converter is one of several things that produce that measurement. An exhaust leak, a drifting oxygen sensor, or an engine running rich will produce it too.
That distinction is why every repair on this site is written as conditional on a specific test result. Replacing a converter that was never the problem is a four-figure mistake, and it is one people make constantly because a website told them what their code "means".
When we are wrong
We will be. Automotive diagnosis is deep and vehicles differ, and no reference site of any size is free of errors. When you find one, tell us — we correct it and we say what we changed.