Security Verification

What can be independently checked, how the encryption works, and where the architecture is open to inspection.

What Can Be Independently Verified

Security claims mean nothing without proof. Here is what you can check right now, without taking our word for anything.

What	How to verify	What it proves
TLS 1.3	Open your browser tools on comsend.com.au. The TLS certificate is issued by Let's Encrypt with a publicly auditable certificate transparency log.	All traffic between your browser and Comsend is encrypted. The certificate is independently issued and logged.
DMARC at reject	Run dig TXT _dmarc.comsend.com.au. The policy is p=reject — the strictest setting.	Nobody can forge email from comsend.com.au. Attempts are rejected, not just warned about.
DNS integrity (DNSSEC)	Check the domain at dnsviz.net or run dig +dnssec comsend.com.au.	DNS responses are cryptographically signed. They cannot be spoofed in transit.
Signal Protocol	Comsend links to Signal — an open-source, independently audited messaging protocol. Signal's cryptography has been reviewed by academic security researchers and is documented at signal.org/docs.	Messages are end-to-end encrypted using a protocol that has survived independent academic scrutiny — not a proprietary scheme.
APK signing	Download the APK and run apksigner verify comsend.apk. The signing certificate is unique to Comsend.	The APK has not been tampered with since it was built. The signing key proves it came from us.
Infrastructure (partial)	WHOIS on our IP ranges, TLS certificate transparency logs, and DNS records are all publicly visible. Our DMARC aggregate reports are sent to a publicly routable address.	The existence and ownership of the infrastructure can be verified. The network is real; the domain controls are real.

End-to-End Encryption

The question "does Comsend use end-to-end encryption" has a specific answer — and it depends on which channel you are asking about.

Signal messages (text, photo, document)

Yes — end-to-end encrypted by the Signal Protocol. The message content is encrypted on the sender's device and can only be decrypted by the recipient. Comsend cannot read these messages in transit.

Retention is conditional: the message is only retained if the recipient is a client with an active matter. When a message arrives, the system matches the sender's phone number against the firm's client records. If a match is found, the decrypted message is captured, hashed, and transmitted to the vault over TLS 1.3 — filed to the client matter and protected by legal professional privilege. If no client match exists, the message is not retained. It remains only on the devices of the participants, as Signal intended. The system does not retain communications that fall outside the lawyer-client relationship.

SIP calls

TLS 1.3 between the app and the phone system. The audio is encrypted from the handset to your PBX. The PSTN leg — from your telephony provider to the recipient — is standard telephony. This is the same security model as every business phone system in Australia: encrypted to the exchange, then carried on the public network. If you require end-to-end encrypted voice, Signal calls (which Comsend does not yet relay) or a separate E2EE voice channel would be appropriate.

Push notifications

TLS-encrypted to your own server. The push payload contains a wake token — never message content. The app fetches the actual message directly from your server after receiving the wake signal. No third party sees the content.

"Security Through Obscurity"

The criticism — that a private system cannot be secure without public auditing — deserves an honest answer.

Comsend is not a mass-market consumer app. It is deployed on a per-firm basis on infrastructure the firm controls. The security model is not "trust us because you cannot see the code." It is:

For messaging: Signal Protocol — publicly audited, academically reviewed, open-source. We did not build the encryption. We built the conditional retention that sits behind it — lawyer-to-client only, privilege-protected, client-matched.
For transport: TLS 1.3 — the same protocol your bank uses. Publicly specified, independently implemented, verifiable through certificate transparency.
For the APK: Signed with a unique key. You can verify it has not been tampered with since we built it.
For DNS: DNSSEC-signed. You can verify that the records you receive are the records we published.

What we do not publish — and will not publish — is the source code of the vault ingestion engine, the disclosure gateway, or the infrastructure orchestration. These are proprietary legal practice tools, not cryptographic primitives. They do not need to be open-source to be secure — they need to be verifiable in operation. The evidence they produce must survive a courtroom challenge, which is a higher standard than a GitHub repository.

On audits: The platform undergoes authorised penetration testing by our internal Hatter's Tea Party unit — the same team that provides security testing to client firms. Results are available to firms under NDA. A public audit report would describe a single deployment at a single point in time; the architecture is per-firm, and each firm's deployment is independently verifiable by their own IT security team or external auditor.

Physical & Infrastructure Security

Comsend runs on hardware the firm controls — typically a dedicated server or a private cloud node in an Australian data centre. Physical access is controlled by the firm, not by Comsend.

Access logs: The endpoint monitoring agents on every container log all access — including physical console access, SSH sessions, and API calls. These logs are retained and auditable by the firm.
Network segmentation: The phone system, mail server, edge server, push server, and vault run in separate containers on separate VLANs. A compromise of the web-facing edge server cannot reach the vault.
Patch management: All containers run current LTS operating systems. The vulnerability scanner flags unpatched software within hours of a CVE being published.

The Bottom Line

If your threat model is a foreign government intercepting your communications, Signal's end-to-end encryption — which Comsend uses — is the appropriate defence. Signal has been independently audited; its protocol is public; its security is mathematically provable.

If your threat model is losing a client message when a phone is upgraded, failing to disclose a document because nobody filed it, or being unable to prove to a court that a communication happened — then Signal alone is not enough. You need retention, filing, and chain of custody — which is what Comsend adds. But only for communications within the lawyer-client relationship. If the sender is not a client, nothing is retained. The system applies retention only where privilege applies.

Comsend does not replace Signal's encryption. It sits behind it — capturing what the lawyer already received, and making it survivable as evidence. The two systems solve different problems. Together, they solve both.