01
Machine message
A sensor publishes a raw value over MQTT.
Built for regulated & quality-critical systems
Make machine messages understandable, comparable and reviewable
Standard brokers move MQTT messages. TrailMQ adds context, comparison and audit evidence.
Control MQTT access, explain broker decisions, enrich machine messages with domain context, and turn live values into reviewable evidence — without changing your machines or your broker.
Open source on GitHub
Self-hosted via Docker
No cloud dependency
No vendor lock-in
topicproduction/line1/filler/
batch/4711/temperature
batch/4711/temperature
contextline1 · filler · batch 4711
live28.4 °C vs 22.5 °C baseline
decisionaccepted → enriched → compared
deviation+26.2% critical
policypolicy@v7 active
prev a91f…7c2 → this 3e08…d14
The problem
Fast transport is easy.
Fast transport is easy.
Explaining it later is hard.
Months later, teams need to explain who changed what, which policy was active, what a machine message meant, and why something was allowed, blocked, deferred or calculated. A standard broker only tells you the message moved. TrailMQ keeps that evidence structured and reviewable from the moment it happens.
02
Add context
Line, machine, batch and metric extracted from topic and payload.
03
Compare & decide
Live value checked against historical baselines and policy.
04
Baseline resolved
Expected reference values recorded with the decision.
05
Audit evidence
Everything linked into one reviewable, tamper-evident record.
You keep MQTT. TrailMQ adds a control and evidence layer on top — no change to your machines, no change to your broker, no proprietary lock-in. Why MQTT alone is not GxP compliant
Capabilities
Not just transport.
Not just transport.
Contextual, controlled messaging.
TrailMQ does not replace your broker. It sits in front of it and adds control, domain context, historical comparison and reviewable evidence.
Policy-controlled topic access
Define who can publish or subscribe to which topics, under which conditions. Policies are versioned and always part of the audit record — so you can prove not just what happened, but which rule allowed or blocked it.
Immutable audit trail
Every decision is written to a structured, append-only log that stays readable and reviewable later. No post-processing required.
Identity & change tracking
Know which client, user or service triggered a message, when a configuration changed and who changed it. Identity is connected to every event.
Structured, readable evidence
Not raw logs. Not binary blobs. Structured records a validation engineer, auditor, QA team or OT owner can review without a developer in the room.
Domain context enrichment
Turn technical MQTT topics into readable machine, batch and metric context — making audit records understandable for QA, OT and engineering teams.
Live vs. historical comparison
Compare live MQTT values against REST-fed historical baselines and record KPI deviations as audit-linked evidence.
Who it's for
Built for teams that need to explain later
The layer between your MQTT traffic and the reviewable evidence an audit, a deviation or a regulator will eventually ask for.
OT & Production teams
Know who changed a setpoint, when it happened, whether the right policy was active and what the message meant in machine, batch or metric context — without touching your PLCs or sensors.
Pharma & Life Sciences
You operate under GMP, and auditors ask questions months later. Get structured, reviewable evidence for message decisions, deviations and technical controls — ready when you need it.
IIoT Platform engineers
You build the infrastructure others rely on. TrailMQ runs as a Docker-based Starter Kit, integrates without code changes, and produces evidence that downstream systems can consume.
Standard broker vs. TrailMQ
The questions a broker can't answer
A standard broker tells you messages moved. Here is what it leaves unanswered.
Question you need to answer
Standard broker
With TrailMQ
Who changed this topic configuration?
No record
Identity tracked
Which policy was active when this message was sent?
Not tracked
Policy version in audit log
Why was this message blocked?
No explanation
Decision recorded with reason
What does this MQTT value mean in process context?
Topic string only
Machine, batch & metric context
Was this live value normal or deviating?
Not available
Compared with historical baseline
Can I prove this data was not modified in transit?
No
Tamper-evident log
Is this evidence ready for a GMP audit?
No
Structured, reviewable records
TrailMQ does not compete with your broker. It makes broker decisions, machine context and deviation evidence visible, reviewable and explainable.
Regulated environments
Designed for where traceability is not optional
TrailMQ serves as a technical control in support of compliance processes. It does not replace full regulatory assessment or validation — it gives you the evidence those processes depend on.
- Immutable audit evidenceHash-chained, tamper-evident trails
- Segregation of dutiesUser and role-based access controls
- Data integrity controlsMessage-level verification
- Supports validation workflowsIQ / OQ / PQ documentation ready
- GAMP alignmentTopic-level permissions across validation phases
GMPGood Manufacturing Practice
GAMP 5Risk-based validation
21 CFR Part 11Electronic records
Data integrityALCOA+ principles
TrailMQ is early-stage software and a technical control, not a compliance guarantee. Validation, risk assessment and procedural controls remain the responsibility of the regulated organization.
Worked example
From a live value to reviewable deviation
A machine publishes a live temperature. TrailMQ extracts the domain context, resolves a historical baseline, and records the deviation as evidence.
1
Feed a historical baseline through REST
An external system provides a baseline for line1 / filler / temperature.
// expected reference for this context
{
"context_key": { "line": "line1", "machine": "filler", "metric": "temperature" },
"baseline": { "value": 22.5, "unit": "C", "source": "historical_average_30d" },
"limits": { "warning_percent": 10, "critical_percent": 20 }
}2
Receive a live MQTT value
A machine publishes 28.4 °C to production/line1/filler/batch/4711/temperature.
3
Produce audit-linked deviation evidence
TrailMQ links the live value, context, baseline, deviation and decision trace into one reviewable path.
{
"metric": "temperature",
"live_value": 28.4,
"historical_value": 22.5,
"deviation_percent": 26.22,
"severity": "critical",
"context": { "line": "line1", "machine": "filler", "batch_id": "4711" }
}If required historical context is missing, TrailMQ does not silently skip the calculation. Missing context becomes explicit: deferred, retryable and reviewable.
Trusted Industrial AI
A trustworthy foundation before the model
In GMP and quality-critical environments, the question is not only whether an AI model is powerful — it's whether the data, context and decisions around that model can be trusted and reviewed.
AI needs more than raw machine data
Before a model can reason about industrial data, the value needs domain context: machine, line, batch, metric, policy state and process relevance.
Deterministic context first
Extract domain context, resolve historical baselines, calculate deviations and link the result to audit evidence — deterministic by default.
Trust comes from traceability
If AI later suggests or explains an event, TrailMQ provides the surrounding evidence: what data was used, which context applied, which baseline resolved and which decision path was recorded.
04 Trusted Industrial AI
03 TrailMQ evidence layer
02 MQTT broker
01 Machines & sensors
TrailMQ is not positioned as an AI model or analytics platform. It is the controlled MQTT evidence layer that trusted Industrial AI can build on.
Quickstart
Start TrailMQ with one command
Clone the repo, run the guided launcher and choose your first Starter Kit. TrailMQ prepares the stack for you.
The guided launcher prepares runtime folders, creates local demo certificates when needed, generates evaluation credentials and starts the selected recipe.
After launch, use the Web UI or the REST API to inspect topics, resolve policies, review queues and validate audit evidence.
No migration. No downtime. No code changes on your MQTT clients.
TrailMQ is early-stage software. Evaluate it carefully before using it in critical production environments. Free to evaluate — contact us for production licensing.
# Clone the TrailMQ deployment repo
git clone https://github.com/RainerGewalt/TrailMQ.git
cd TrailMQ
# Start the guided launcher
./trailmq launch# Choose your Starter Kit
[1] Secure MQTT Core
✓ Runtime folders prepared
✓ Config ready
✓ Evaluation credentials generated
✓ Active recipe set
✓ Stack is upWeb UI http://localhost/trailmq/
REST API http://localhost/api/v1
MQTT TLS localhost:8883
MQTT WS ws://localhost/mqtt
FAQ
Can MQTT be used in GxP environments?
MQTT can be used in regulated manufacturing, but a broker alone is not enough. The real question is whether message decisions can be controlled, explained and reviewed later.
Can an MQTT broker be GxP compliant?
MQTT itself is not GxP compliant or non-compliant — it is a messaging protocol. Compliance depends on the surrounding controls: identity, access policies, audit trails, data integrity, change control and validation evidence.
What about data integrity and audit trails?
Standard brokers move messages. They usually do not explain who changed what, which policy was active, why a message was allowed or blocked, and whether the evidence is complete enough for later review.
Does TrailMQ make a system automatically compliant?
No. TrailMQ supports compliance by generating technical evidence. Validation, risk assessment and procedural controls remain the responsibility of the regulated organization.
Does TrailMQ provide a REST API?
Yes. TrailMQ exposes product functions through a REST API so teams can inspect topics, resolve policies, review queues, validate audit evidence and integrate TrailMQ with scripts, local checks, CI pipelines, monitoring tools or external systems.
Can TrailMQ compare live values with historical baselines?
This is part of the planned plugin layer. Historical baselines can be provided through REST, live MQTT values can be enriched with domain context, and KPI Lite can calculate deviation metrics linked to audit evidence.
Standard brokers move MQTT messages.
TrailMQ makes machine communication understandable, comparable and reviewable.