Skip to content

Find where a value went

Given an identifier, such as an account number or an email address, complydoc lists every document it was read from, every step of a pipeline that still held it, and every host it was sent to, across all the runs in a folder.

Terminal window
complydoc find "GB82 WEST 1234 5698 7654 32"
IBAN •••• •••• •••• •••• ••54 32 in 3 runs of 3 reports
contracts-ingest pipeline, 2026-10-13 21:34 e8dd1c8
in employee-record.pdf page 1
1 DirectoryLoader load, 1 time
2 StripPathMetadata transform, 1 time
3 RecursiveCharacterTextSplitter split, 1 time
4 InMemoryVectorStore store, 1 time → localhost
sent to localhost

Left out, the value is asked for without being shown, which also keeps it out of your shell history. --in names report files or folders to search, and defaults to .complydoc. --json prints the result as JSON. In the viewer, the Find a value page asks the same question of every run it serves.

Reports do not hold the values they found, only a fingerprint of each one: a hash of its category and its characters. The value you give is hashed the same way and the reports are searched for the hash. So the question can be asked of reports that never held the value, and the value asked about is written nowhere.

For a pipeline recorded with cd.observe, every step is listed, the ones that never held the value too: where it stopped is as much the answer as where it went. A run that masked identifiers shows the step after which the value was gone.

A value is found only where a run looked for its kind and read it as it is written here. Spacing does not matter. A finding’s fingerprint can be given in place of the value.

A fingerprint is not a secret: a value with few possibilities can be recovered from one by trying them all.

From Python:

import complydoc as cd
trail = cd.find_value("GB82 WEST 1234 5698 7654 32", ".complydoc")
for run in trail.runs:
print(run.name, run.sent_to)