COMPARISON DOSSIER · ISSUE 2026

Torzon vs the Alternatives 2026: Comparing Onion Uptime, PGP and Escrow Claims

Few people really ask which onion market is best. They ask whether the one in front of them is real, and whether something else would be safer. This file sets Torzon beside a handful of names that keep circulating in 2026 and runs every one of them through the same short set of questions. We sign a single address, our own, and we are blunt about the rest.

Published 2026-08-22 · by Mira Kessel, Field Notes desk

GROUND RULESwhere it stops

How this comparison stays fair

A comparison earns its keep only when every entry meets the same test. So we do not grade on looks, on how busy a forum thread feels, or on a vendor count that nobody can audit. We put four plain questions to each market and answer them the same way each time. Where a thing is not confirmed, the cell reads pending or unverified rather than a guess dressed up as fact.

One rule sits above the others. We publish and sign one onion address, the Torzon reference on our front page. For any other market we can print an address so you can orient yourself, but we have not signed it. Read those as a lead to confirm at that market's own source, never as a nod from us.

METHODsame test, each name

One yardstick, held against each market

The trick to a fair table is boring on purpose. We do not invent a fresh test per market. The same four questions meet Torzon and every alternative, and only then do the answers line up next to each other.

One yardstick across marketstorzonnexusvortexYardstickLike for like
One rubricEscrow, PGP signing, uptime honesty, and phishing surface. Four questions, no bonus points for a slick page.
Applied in turnTorzon and each alternative meet the very same four. Nothing gets an easier version of the test.
A like-for-like readThe output is a comparison you can check, not a ranking and not a recommendation to buy.
THE FIELDnames in the frame

The markets in this comparison

Here are the names, with the address we hold for each. Only the Torzon row is ours to sign. The others carry an onion for orientation and a plain unverified tag, because we have not put our signature behind them.

Addresses shown for reference. Torzon is our signed canon; confirm every other market at its own source before you act.
MarketOnion URLWhat we holdAction
Torzonhttp://torzonguqmlfy2kfi5tjbnt4bp3idtkjzi4qtupmhpdihjftomjtdzqd.onionSigned canon
Nexushttp://nexusakrv5pwd5jtzkycsmp3sllliq6nxilutfug2o3rlmdxmqlzz2ad.onionUnverified
Vortexhttp://vortexqttelw7m627vd25uhzielmoqbbbep6xn3hoa6re2xiac3tomad.onionUnverified
Marshttp://marskp4ozu3nv2ez3in5ofyukovali7o5ioxyuvubeus74cu2bjl5nid.onionUnverified

A tag here is about our own checking, not a safety score. An unverified market is not judged unsafe; it is simply one we have not signed.

Nexus, briefly

Nexus carries an onion above for orientation only — we have not built a signed canon for it, so treat the address as a starting point to verify at Nexus's own reference, not as something this page vouches for.

Vortex, briefly

Same limit applies: Vortex's address is listed for reference, unverified by this site. If you're comparing Torzon against Vortex specifically, do that comparison's homework at Vortex's own signed source, not here.

Mars, briefly

Mars is included because it's one of the names readers ask about most alongside Torzon. As with the others, we hold no signature for it — mars-darknet.com runs its own dedicated verification reference if that's the market you're actually trying to confirm.

SIDE BY SIDEquestion by question

Torzon against a typical alternative

This is the whole point of the page. The left column is a question. The middle is what Torzon actually does. The right is the pattern we see across a lot of alternatives, written as a habit rather than as any single market.

The Torzon column reflects what this project runs today. The alternative column describes common patterns, not one named market.
QuestionTorzonA typical alternative
Escrow modelMonero, held in escrow until an order clears.Mixed. Some hold funds, others lean on direct pay that protects the seller more than you.
PGP-signed directoryThe model is in place; the key is not published yet (Phase 0).Often just a pinned forum link or a screenshot, with nothing to verify offline.
Uptime signallingStays on checking until a probe answers. No fixed green light.Frequently shows a hard-coded online that never changes, up or down.
Phishing surfaceLong lookalike onions, so we add a character-by-character link check.The same lookalike risk, with rarely any check offered to the reader.
What the page verifiesOne address, held against our own canon.Usually nothing you can confirm on the page itself.
TWO HABITSwhich one survives

How people pick, and which habit holds up

Picking by the name

The address looks familiar, the logo is right, a thread swears it is current. None of that is hard to fake. A clone copies the look in an afternoon, and the name is the first thing it steals.

Picking by the signature

You compare the full onion against a signed source and confirm the key. It is slower, and it is the only step a copy cannot pass. Everything on our scored review leans on this one habit.

Every market on a list like this looks trustworthy right up until you try to prove it. The one you can verify is the only one you can honestly compare.Field Notes desk
OPEN QUESTIONSwhat we cannot say

What this comparison will not tell you

It will not crown a winner. We have not placed orders on any of these markets, so we cannot speak to a vendor, a price, or how a dispute ends. It will not rank the alternatives against each other either, because we hold no signed address for them and a wrong onion is worse than no onion. And it is a snapshot: a market that answers today can vanish next week. For the one name we do stand behind, read the full Torzon verdict and verify the address yourself.

WHY COMPARE AT ALLthe actual use of this page

What a market comparison like this is actually useful for

A side-by-side comparison of Torzon against named alternatives is not a recommendation to use any of them — it is a way to calibrate expectations. Reading Torzon's practices in isolation makes it hard to tell whether a given habit (a particular escrow model, a particular mirror-rotation cadence) is a strength, a weakness, or simply how every market in this space operates.

What comparison reveals that a single review cannot

A single-market review can describe what Torzon does. Only a comparison shows whether that is unusual. A mirror-rotation pattern that looks concerning in isolation might be standard practice across the field, or it might be a genuine outlier — the comparison is what tells the difference.

COMPARISON FAQplain answers

Questions about this Torzon comparison

Is this comparison sponsored by Torzon or any competitor?

No. No market pays for placement or a favorable reading in this comparison, and the same criteria apply to every name in the field regardless of which one this reference happens to focus on elsewhere.

Why does only Torzon's address get signed on this reference?

This reference maintains its own signed canon for Torzon specifically. It does not operate or verify addresses for the other markets named here, and does not claim to — readers should verify any other market's address through that market's own trusted source.

Does a market scoring well here mean it is safe?

No comparison of this kind can promise safety. It weighs observable operational practices side by side; independent verification before connecting to any of the named markets remains necessary regardless of how the comparison reads.

How often does this comparison get updated?

When a criterion-moving change happens to any named market — not on a fixed schedule unrelated to what actually changed in the field.

READING THE MATRIXhow to use the side-by-side table

How to actually use the side-by-side comparison

A comparison matrix invites a quick glance and a snap judgment — scanning for the most green cells and picking that row. That is the wrong way to read this one.

Weight the rows, do not just count them

Not every criterion carries equal weight for every reader. Someone prioritizing anonymity above all else should weight the escrow-and-payment-privacy row far above a row about interface polish. Someone new to Tor markets entirely might weight ease-of-verification highest. The matrix reports facts; the weighting is a judgment only the reader can make for their own situation.

A single red cell is not disqualifying on its own

Every market in this comparison, including Torzon, has at least one criterion where it does not lead the field. The question worth asking is not "does this market have any weakness" — every one does — but whether that specific weakness matters for what a given reader actually needs from a marketplace.

The matrix ages faster than the method behind it

Any individual cell in this table can go stale the moment a market changes its escrow model, rotates a key without warning, or has a mirror outage this page has not yet caught. The method — checking the same criteria the same way across every name — stays valid even when a specific reading needs a refresh.

SWITCHING COSTSwhat a name change costs a buyer

What it actually costs to switch markets after choosing one

A comparison like this can make switching sound frictionless — a row loses ground, so the reader moves to the row above it. In practice, an established account on one market carries a vendor relationship history, an accumulated escrow trust level in some cases, and a familiarity with that specific market's verification habits that does not transfer anywhere.

Verification habits do not transfer automatically

Someone comfortable checking Torzon's PGP fingerprint and mirror pattern has to relearn the equivalent process for a new market from scratch — a different signing key, a different rotation cadence, a different set of known clone patterns. That relearning period is itself a window of elevated risk, since unfamiliarity is exactly when a phishing clone is hardest to spot.

When switching is still worth it

None of this means staying put is always correct — a market with a documented pattern of extended outages or an unexplained key rotation is a real reason to move. It means switching should be a deliberate response to a specific, checkable concern, not a reflexive reaction to one comparison row looking marginally better elsewhere.

COMMON MISTAKESreading this table wrong

Common mistakes when reading a market comparison like this one

The single biggest mistake is treating any row in this comparison, Torzon's included, as fixed rather than as a snapshot. Uptime, PGP signing status, and phishing exposure all change over a market's lifetime, sometimes within days, and a comparison table built in 2026 will drift from reality faster on the markets that change fastest. Re-check the live status for whichever market you are actually about to use — Torzon or otherwise — rather than treating this page as a permanent verdict.

Picking a "winner" instead of matching the criteria that matter to you

This comparison deliberately avoids naming an overall winner, and reading one into it anyway is a mistake. Torzon scoring well on PGP signing does not make it the right choice for someone who cares most about vendor selection in a specific category, and a competitor scoring well on uptime does not make it safer if its PGP practices are weaker. Match the specific row that matters for your situation instead of collapsing the whole table into a single ranking it was not built to produce.

Assuming a market absent from this table was rejected

A market not appearing in this Torzon comparison is not a statement that it failed some cut — it usually just means this file has not been extended to cover it yet, or that there was not enough independently checkable information about it to score fairly. Absence from this table is neutral, not a red flag, and should not be read as an implicit warning about a market this page simply has not addressed.

ADDRESSES

Everything worth copying sits one click away on the home page: the signed address, the character check, and the key. Pull them from there, never from a table built only for reading.

Open the mirror list