ESS PRO · Platform

Every server. Every packet.
Your rules.

AMVLET's Secure Border Gateway gives your organisation complete control over who can connect to your Matrix deployment — keeping approved partners seamlessly connected while blocking everything else at the door.

Get started → Talk to sales →
Application-level inspection Zero unauthorised connections Pluggable rule engine Full audit trail
100%
Of traffic inspected at the application layer — every request checked before it reaches your server
0
Unauthorised connections allowed through — the gateway is the only entry point
2
Core control modes: server federation and client access — covering every connection type
Custom rules your organisation can stack — no limit on the security policies you layer
Live preview

One gateway. Total visibility.

Every connection attempt — from partner servers, external clients, or unknown sources — passes through the gateway and is evaluated against your rules in real time.

AMVLET · Secure Border Gateway
Active
Decision Origin server Type Latency
ALLOWED element-hq.matrix.org Federation 8 ms
ALLOWED bnd.bund.de Federation 14 ms
BLOCKED unknown-relay.net Federation
ALLOWED regulator.finance.gov.uk Federation 22 ms
BLOCKED 192.168.99.1:8448 Client
ALLOWED partner.amvlet.cloud Federation 11 ms
BLOCKED random-matrix-node.xyz Federation
ALLOWED ministry-comms.gouv.fr Federation 18 ms
Active rules
1
Server allowlist
Permit only approved Matrix servers
ON
2
Client validation
Approved apps only
ON
3
Event type filter
Restrict by Matrix event type
ON
4
Audit log
All decisions recorded
ON
Core capabilities

Federated, not out of control.

The Secure Border Gateway gives regulated organisations what open federation alone cannot: the ability to connect selectively, enforce precisely, and audit completely.

Federation Control

Decide precisely which Matrix servers can communicate with yours. Maintain a list of approved partner organisations and the gateway will silently reject all others — no manual intervention required. You stay connected to who matters; you remain invisible to everyone else.

Client Restrictions

Ensure only approved Matrix applications can connect to your homeserver. Block personal and unknown clients automatically, enforce corporate-only access, and eliminate shadow IT — without disrupting any legitimate user on an approved device.

Pluggable Rule Engine

Build your security policy as a pipeline of rules, each one adding a precise layer of control. Rules can inspect any part of a request — server identity, user ID, event type, destination room. New rules can be deployed without any service interruption.

How it works

Five layers of protection.

Unlike a conventional firewall, the Secure Border Gateway understands the Matrix protocol. Here is what that means in practice.

1

Application-level inspection

A traditional firewall sees only IP addresses and port numbers — the equivalent of knowing that a car arrived at your gate, but nothing about who is inside or what they are carrying. The Secure Border Gateway reads and understands the Matrix protocol itself. It knows whether a request is trying to join a room, send a message, or establish server-to-server federation, and it can apply different rules to each scenario.

2

Full metadata enforcement

Security decisions are made using the complete context of every request: which Matrix server it came from, the Matrix ID of the requesting user, the type of event being sent, the destination room, and the precise timestamp. No request can disguise itself by changing a single attribute — the gateway evaluates the full picture before making any decision.

3

Pipeline-based rule processing

Requests pass through a sequence of security checks, one after another. Each rule can allow, block, or modify the request before passing it to the next rule in the pipeline. This layered approach means complex security policies can be built from simple, composable building blocks — and individual rules can be updated without affecting the rest of the pipeline.

4

Custom rule development

Organisations can write and deploy their own rules to meet specific compliance obligations. Whether you need to enforce regional data residency restrictions, apply sector-specific communication policies, or integrate with an existing identity management system, the pluggable architecture accommodates it — without modifying the core gateway or requiring downtime.

5

Zero-trust by default

The gateway operates on the principle that no connection is trusted until it has been explicitly approved. Every request — whether from a known partner server, an internal client, or an external federation peer — must satisfy the applicable rule pipeline before being passed through. There is no implicit trust, no bypass, and no exception that is not written into your policy.

The Matrix federation challenge

Why the most open communications protocol needs the most precise controls.

Matrix is built on a powerful idea: every organisation owns its own server, and servers can talk to each other directly — no single company in the middle, no central point of control, no vendor holding your data. This is what makes Matrix genuinely sovereign. But for regulated organisations, sovereignty without control is not enough.

The problem with open federation

When a Matrix server joins the global network without any gatekeeping, it becomes reachable from any other Matrix server on Earth. For a commercial collaboration platform, that openness is a feature. For a government ministry, a defence contractor, or a financial institution, it represents a compliance boundary that cannot be left undefined.

Consider the practical reality: a defence organisation needs to share classified project rooms with specific cleared contractors, but not with every other Matrix user in the world. A hospital network needs to exchange patient-related communications within an approved care pathway, but must be able to demonstrate that no message ever crossed an unapproved boundary. A financial services firm must enable encrypted communications with its regulator — but MiFID II requires that those communications be monitored and archived to a defined standard.

Traditional firewalls cannot solve this problem. They work at the network layer, seeing only IP addresses and port numbers. Matrix federation runs over standard HTTPS, so a conventional firewall cannot distinguish between a legitimate partner server and an unknown one attempting unauthorised access. You need something that understands Matrix itself.

Defence & Government
Federate only with cleared partners and approved agencies — remain invisible to all others
Financial Services
Open a controlled channel with regulators and auditors — with a complete audit trail of every message
Healthcare
Restrict clinical communications to approved care pathways — patient data never crosses an unapproved boundary

How the gateway works — in plain terms

Think of the Secure Border Gateway as the security desk at the entrance of a secure building. A regular firewall is the front door — it checks whether you have the right key. The SBG is a trained security officer who not only checks your credentials, but knows exactly which rooms you are authorised to enter, what you are permitted to carry, and keeps a detailed log of your visit. Nothing gets through that the policy does not explicitly allow.

In technical terms: when any external server or client tries to connect to your AMVLET homeserver, that request first arrives at the Secure Border Gateway — not at the homeserver itself. The gateway evaluates the request against your configured rule pipeline. If the request passes all rules, it is forwarded to your homeserver. If it fails any rule, it is blocked instantly and the decision is logged. Your homeserver never sees traffic it was not supposed to receive.

Because the gateway understands the Matrix protocol at the application layer, it can see context that no IP firewall could access: which Matrix server is sending the request, which user is making it, what type of Matrix event they are attempting to send, and which room they are targeting. That rich context is what enables policy-grade enforcement — not just "block this IP address" but "allow this specific server to federate, but only for these event types, only to these rooms."

Federated, not out of control

The key insight — the one that changes how regulated organisations think about Matrix — is that federation and control are not opposites. The question is not whether to open your server to the world or close it completely. The question is whether you have the tools to make a precise, auditable, enforceable decision about every connection attempt.

With the Secure Border Gateway, you do. You can federate with the five partner organisations your compliance framework requires, while remaining unreachable by every other Matrix server on Earth. You can allow your regulator to join specific audit rooms, while maintaining a complete record of every message that crossed that boundary. You can connect your overseas offices into a single sovereign deployment, while enforcing the data residency requirements that apply to each jurisdiction.

This is sovereign federation. Not a binary choice between open and closed, but a precise, configurable, auditable boundary that you define and control completely.

Questions

Everything you need to know.

The questions security, compliance, and engineering teams ask before deploying a Secure Border Gateway.

What is a Secure Border Gateway in plain terms?
It is a security checkpoint that sits between your AMVLET server and the outside world. Every connection request — whether from a partner organisation's Matrix server, a mobile client, or any other source — must pass through the gateway before it can reach your deployment. The gateway checks the request against your rules. If it is allowed, it goes through. If it is not, it is blocked and logged. Nothing reaches your server without passing this check, and nothing that passes it does so without a record being created.
How is this different from a regular firewall?
A conventional firewall works at the network layer — it can block or allow traffic based on IP addresses and ports, but it cannot look inside the traffic to understand what it actually is. The Secure Border Gateway works at the application layer. It understands the Matrix protocol, which means it can make decisions based on who is sending the request, what type of Matrix event they are sending, which room they are trying to access, and which server they are federating from. This depth of inspection is not possible with any IP-layer firewall — and it is precisely what regulated environments require.
Can I allow specific partner organisations while blocking everyone else?
Yes — this is the primary use case. You configure an allowlist of approved Matrix servers, and the gateway permits federation only with those servers. All other federation requests are rejected automatically, without any manual intervention. You can also configure different rules for different partners: one organisation might be permitted to send any event type to any room, while another is restricted to specific rooms or specific event types. The level of granularity is entirely up to you.
What happens to blocked requests?
Blocked requests never reach your homeserver — they are rejected at the gateway before any content is processed. The block decision is recorded in the audit log, including the identity of the requesting server, the type of request, the specific rule that triggered the block, and the precise timestamp. This log is available for compliance reporting, security investigations, and regulatory audit purposes. Blocked servers receive a standard rejection response and have no visibility into your deployment's internal structure.
Does the gateway affect message delivery speed?
For approved traffic, the performance impact is negligible. Rule evaluation is designed to be fast — most decisions complete in single-digit milliseconds. In normal operation, the gateway adds no perceptible latency to legitimate communication. For blocked traffic, the performance difference is irrelevant: those requests are stopped before they consume any significant server resources, which actually improves overall performance compared to a deployment that processes every connection attempt regardless of its legitimacy.
Can rules be updated without taking the gateway offline?
Yes. The Secure Border Gateway supports live rule updates. You can add, modify, or remove rules from the pipeline and the changes take effect immediately — without restarting the gateway or interrupting active connections. This allows your security team to respond to new threats or policy changes without any service disruption. New rules can be tested in an audit-only mode before being promoted to enforcement, which means you can validate the impact of a policy change before it affects live traffic.

Federated on your terms.

See how AMVLET's Secure Border Gateway gives your regulated organisation the precise federation control it requires — without sacrificing interoperability.

Start free trial → Talk to our team →