Separate the things you know
Create a private worksheet with three headings: known facts, remembered possibilities and unsupported guesses. “The file was exported from this application” and “I probably used a year in the password” do not have the same evidential weight. Start with the strongest information.
Define one testable candidate family
An invented example has two spellings, three year suffixes and two punctuation endings. That family contains at most twelve combinations before removing duplicates. Adding an unconstrained character slot multiplies the work again. Small, explicit rules make it easier to detect omitted or repeated candidates.
A search has two components
Candidate generation proposes strings. A wallet-specific verifier checks them against the appropriate encrypted data and verification rules. Fast generation does not imply fast verification. Measure performance only in the intended environment with an appropriate disposable fixture, and label any timing as specific to that test.
Define the stop and interpretation rules
- Record the exact candidate family and its unique count.
- Set a reasonable time or candidate budget before running a test.
- Keep a record of completed and interrupted runs.
- Treat “no match” as applying only to that candidate family and verifier setup.
- Reassess the backup format and assumptions before expanding a search by many orders of magnitude.
What recovery cannot promise
An Ethereum address alone does not reveal the signing private key. Strong, unknown encryption passwords are not made recoverable by a website or a larger word list. A plausible assessment depends on what survives: the relevant backup, partial memories, correct software identification and a way to verify a result.
Prepare the first enquiry
Send the application name, approximate creation date, broad backup type and what you have already tried. Do not include the actual candidate list in an initial email. Use mail@keychainx.io to discuss what evidence is appropriate for a private assessment.