Hashcat refused to load your hash file and stopped at No hashes loaded. Nothing ran, no keyspace was touched — the job died before it started. The line printed above that message is the actual diagnosis: Token length exception, No hash mode matches or Salt-value exception each point at a different problem.
Five causes, in the order they actually occur. All three sub-errors plus the bare "says no hashes loaded" variant are covered.
Triage: which line prints before "No hashes loaded"?
| What you see | What it means | Go to |
|---|---|---|
| Token length exception | The line is not a valid hash of any kind — wrong format, wrong field count, hidden characters | Causes 1–2 |
| No hash mode matches | Hashcat cannot map your line to the -m mode you selected (or to any mode) | Causes 2–3 |
| Salt-value exception | The structure parsed, but a salt field is empty or zero-length where the format requires one | Cause 3 |
| Just No hashes loaded., nothing else | The file was empty or unreadable — every line was skipped | Causes 4–5 |
| Hashes loaded, but you are here for the John version of this error | Same family, different tool | John: "No password hashes loaded" |
1. The hash line is malformed (Token length exception)
The most common cause overall: the extracted hash is not the full hash. Extraction tools emit long structured lines ($pkzip2$, $7z$, $RAR5$, $office$…), and copying through a terminal, chat window or web form truncates them, wraps them, or strips characters. Hashcat reads the first line, decides it is not any known format, and refuses the whole file.
- Compare the first hash line against the matching entry in hashcat's example_hashes wiki page — same field count, same separators, same length order.
- Copy the extractor output directly from the file it was written to — not from a scrolled terminal pane.
- One hash per line, no quotes, no blank leading line.
This failure has its own full treatment — format checking, line-ending issues and the per-format gotchas: Hashcat Token Length Exception: the full fix.
2. Wrong or missing hash mode (No hash mode matches)
No hash mode matches means the line looks like a hash, but not like a hash of the type -m selected — or like any type hashcat knows. Three sub-cases:
- The -m number is wrong for this file. A PDF extracted with pdf2john is mode 10400/10500/10600 depending on revision; a ZIP is 13600 (WinZip) or 17200/17210 (PKZIP); 7-Zip is 11600. The extractor's name and the hash's $…$ prefix must match the mode.
- You copied a different artifact than you think. A base64 blob from a config file, a bcrypt string from a database dump that is actually truncated, a hex digest with the algorithm prefix stripped.
- The format is a variant hashcat rejects. Some tools emit "john-style" lines; most work in hashcat, a few (older Office, odd Kerberos etypes) need the original extractor re-run.
Identify before you guess: run the line through the hash inspector (offline, in your browser) or hashcat's own --identify, then set -m to the mode the example_hashes page shows for that prefix. If nothing identifies it, the extraction is the problem — go back to cause 3/5.
3. Empty or missing salt (Salt-value exception)
The line parsed as a known format, but a field the format requires — the salt — is empty or zero-length. In practice this means one of three things:
- The extractor produced a broken line. A truncated copy (cause 1) can cut exactly the salt field. This is the same disease with a different symptom.
- The extractor is older than the file format. The classic report on hashcat.net (2017) was exactly this: Salt-value exception on a 7z hash, fixed by upgrading hashcat and the extraction script together. If the source file comes from newer software, update both ends.
- The hash fields disagree with each other. A 2020 report traced the same error to inconsistent length fields inside the 7z hash line — the extractor computed a data length that did not match what it wrote. Re-extract; do not hand-edit the fields.
Rule of thumb: a salt-value exception is almost never "the file has no salt". It is almost always "this line is a damaged or outdated extraction".
4. The hash file is empty, filtered, or unreadable
When only No hashes loaded. prints, hashcat skipped every line:
- File path typo — the file hashcat opened is not the file you edited (cat hashes.txt before launching).
- Encoding damage: UTF-16 output from PowerShell (Format-Hex will show the interleaved zeros), a UTF-8 BOM on line one, CRLF-joined double lines.
- Everything commented or blank after a bad paste.
5. The extraction itself failed
Last resort: regenerate the hash from the original artifact with the current extraction tool — zip2john, 7z2hashcat, pdf2john, rar2john — from the same release as hashcat where possible. If the fresh output loads, the old file was the problem; if it does not, the artifact may not be what it claims to be (a renamed archive, a zero-byte download, a repaired file that no longer contains the encrypted stream you are targeting). And if the goal is understanding the search itself rather than one file: the password cracking time table and how entropy decides the outcome.
Confirm the fix
| Command | Expected result |
|---|---|
| hashcat -m <mode> hash.txt --example-hashes | head | The example for your mode structurally matches your line |
| Launch with a tiny wordlist | 1 hash (or the real count) appears in the status line — loaded, not skipped |
| hashcat --identify hash.txt | At least one candidate mode listed |
Frequently asked questions
What does "No hashes loaded" mean in hashcat?
Hashcat parsed your input and accepted zero lines, so the attack never started. It is a load-time rejection, not an attack result — the line above it (token length, hash mode, or salt-value) says why.
Why does hashcat say "No hash mode matches"?
The hash line does not structurally match the type selected with -m — wrong mode for the extractor you used, or a damaged line that matches no mode at all. Identify the hash first (inspector or --identify), then set -m to the documented mode for that prefix.
What is a salt-value exception?
A structurally recognised hash whose salt field is empty or zero-length. It almost always means the extracted line is truncated or produced by an outdated extractor — re-extract with current tools rather than editing fields by hand.
Where do I find the right hash mode number?
Hashcat's example hashes page maps every $…$ prefix to a mode number; the hash inspector suggests candidates for a pasted line; and the extractor's documentation names its mode. The three should agree — if they do not, the extraction is suspect.
Is this the same as John's "No password hashes loaded"?
Same family — both tools reject the file before attacking it — but the diagnostics differ. John's version of this error is covered separately; the extraction lessons overlap almost completely.