security ↓

Built like a vault.

Your funds never touch our servers. Your trades stay private. Our code is audited, open-source, and watched 24/7. Here's exactly how we keep it that way.

LOCKED smart contract funds live here, not on our servers ↑

Four things we never compromise.

LOCKED on-chain
i.

Funds never leave the contract.

The instant both parties agree, your funds are locked in a smart contract you can verify on-chain. Not in a hot wallet. Not on our servers. Not anywhere we can touch.

  • Non-custodial by design — Middn has no withdrawal key
  • Atomic settlement — both sides clear in the same tx, or nobody clears
  • Auto-refund if the counterparty disappears past the timeout
Read the contracts →
only you & them can read ↓
ii.

Privacy is the default.

Every message between traders is encrypted client-side. The contract sees only what it needs to — amounts, conditions, signatures. The rest? Nobody can read it. Not us, not validators, not the chain.

  • End-to-end AES-256 + X25519 (signal-style, forward-secure)
  • ZK proofs verify conditions without revealing the data
  • Metadata stripped from all on-chain interactions
How ZK proofs work →
AUDITED ✓ { ok } every line, multiple times ↓
iii.

Reviewed by multiple teams.

We don't just say our code is safe — we pay the best security firms in crypto to try and prove it isn't. Trail of Bits, Halborn, Code4rena. Every report is public.

  • 2 external audits completed (Mar & Apr 2026)
  • Active Code4rena contest with $250k prize pool
  • Continuous internal fuzzing & formal verification
See all audit reports →
OK OK OK OK someone's always watching ↓
iv.

Monitored 24/7.

Every contract is watched in real time. Anomaly detection, large transfers, unexpected state — anything weird wakes someone up at 3am. We've got 42 CI gates and on-call engineers across 3 time zones.

  • Forta + Tenderly alerts on every state change
  • On-call rotation in EU, US-East, US-West
  • Pre-funded emergency pause + recovery procedures
Our incident playbook →

What it adds up to.

$0 funds lost
since launch
3 independent
audits
$250k active bug
bounty
100% open-source
code

What we protect vs what's on you.

Security only works if everyone knows their job. Here's the honest split.

Middn protects ↓

  • Contract exploits Audited code, formal verification, fuzzing. We assume hostile inputs.
  • Counterparty default Atomic settlement means neither side can run. Funds release together or refund.
  • Server compromise We can't touch your funds. Even if our entire infra is hacked.
  • Network surveillance E2E encryption + ZK proofs. No metadata leaks.
  • Front-end tampering Subresource integrity, IPFS-pinned UI builds, reproducible.
  • Insider abuse No admin keys exist. There's nothing to abuse.

You're responsible for ↓

  • Your wallet keys We can't recover them. If you lose them, your funds are gone forever.
  • Phishing & fake sites Always check the URL. Always sign on a clean device. Bookmark middn.com.
  • What you trade We're a neutral layer. Knowing your counterparty is your call.
  • Reading what you sign Verify amounts, conditions, addresses in your wallet before signing.
  • Off-platform agreements If you negotiated something outside Middn, we can't enforce it.
  • Tax & legal compliance Your jurisdiction, your responsibility. We don't give legal advice.

Our incident playbook.

Even with audits and monitoring, we plan for the day something goes wrong. Here's exactly what happens.

T+0
min

Detection.

Forta + Tenderly alerts fire to PagerDuty. On-call engineer wakes up. We classify severity (P0 = funds at risk, P1 = service down, P2 = degraded).

T+5
min

Triage & freeze.

If P0, we trigger the emergency pause. This is the only privileged action that exists, and it's a one-way switch: it stops new deals but never moves user funds.

T+30
min

Public disclosure.

Status page updated. Twitter post pinned. Discord & Telegram pings. We tell users what we know, what we don't, and what to do (usually: nothing, your funds are safe in the contract).

T+72
hrs

Post-mortem.

Public write-up. Root cause, timeline, what we fixed, what we changed in our process. No PR-speak. Auditors review the fix before we unpause. Same standard as new code.

questions? ↓

Got a security question?

Our security team reads every email. Bugs go via Code4rena or directly to us — we PGP, we pay, we don't sue.