While answering a few troubleshooting posts here last week I kept hitting the same pattern: the error message names something that is not the problem.

Four real ones:

• zsh: command not found: mem, from an AMR++ run. A red herring. $BWA is a Nextflow environment variable that only exists while the pipeline is running, so copying the command out of the log into a shell leaves the variable empty and the shell reads mem as the command name. The actual failure is one level down.

• 'bash\r': No such file or directory sends people looking for a missing file. The file is there. Its name has a carriage return appended, because the script was saved on Windows.

• A NegativeArraySizeException in HaploView looks like a memory problem, so people raise -Xmx at it. It is a 32-bit integer overflow, and no amount of heap touches it.

• Someone set -Xmx128G for Cytoscape and it still crashed. The .vmoptions file needs LF-only line endings; Notepad saves CRLF and the launcher ignores the setting without saying so.

So I started writing them down properly. It is a static page, 42 signatures so far, grouped by area (Nextflow, Snakemake, conda, containers, file formats, R, Java, HPC, shell). You can paste an error and have it matched, or just read the whole catalogue on the page.

https://bielporte10.github.io/bioerrors/

Two things up front, because they are usually the first questions.

There is no LLM in it. It is pattern matching in the page's own JavaScript: no server, no API key, and nothing you paste leaves your browser. That is deliberate. A general chatbot infers something plausible from the error text, which is exactly how you end up being told to add more RAM for an integer overflow. Every entry is either documented behaviour of the tool or was traced by reading the source, the config or the log involved, and each entry says which.

It only knows what is in it. When it does not recognise an error it says so instead of guessing, and gives you a button that opens a pre-filled GitHub issue so it can be traced and added.

Corrections are very welcome. If an entry is wrong or incomplete I would rather hear about it than leave it up.