TN-Proto
DPIA-ready audit trails for agentic AI systems.
Our analysis of common AI adoption roadblocks keeps pointing to the same data problem: access, trust, and evidence break down at the event layer. The initial release is focused on logging for vibe coders, giving builders a practical way to capture what agents did while prototypes become systems. The protocol itself is transaction-based and built for rigorous event evidence across people, systems, agents, and workflows.
View TN-Proto tiersThe recurring roadblock
AI audit trails need more than logs.
The free flow of secure data is the most critical enabler of technology value, and every meaningful agent action needs evidence that can be reviewed later.
AI programs need better models and a dependable data layer. They need reliable, permissioned, inspectable data movement between people, systems, agents, and vendors. Safe data movement is what lets a firm realize the benefits of its technology investments.
Secure infrastructure still matters. In the agentic era, trust also has to be observable at the event level. Agents can read, summarize, call tools, trigger workflows, and create new data trails. The firm needs to know which event happened, who or what caused it, which fields were visible, and which policy governed the exchange.
What we see in workshops
The paradox is practical.
Too locked down
The right people are blocked.
Teams need fast access to the fields, events, and evidence required to resolve exceptions, serve clients, and approve automated workflows.
Too open
Broad systems see too much.
Logs, dashboards, integration pipes, and support tools often expose more data than the task requires, especially once agents start reading across systems.
Too assumed
Trust boundaries are invisible.
Firms trust infrastructure, and agentic systems now require stronger proof of which agent, user, model, vendor, or workflow had the right to see or transmit a specific event.
Protocol response
TN-Proto creates audit-ready event flow.
TN-Proto is a distributed, peer-to-peer transaction event protocol. It was built for systems created by AI, orchestrated by AI, and operated by AI, where every meaningful action needs a secure event, a clear identity, a controlled path for data access, and evidence that can support review.
The protocol and implementation docs live at tn-proto.org. Cyaxios uses the TN-Proto research direction to think through the infrastructure side of agentic AI: what data an agent can see, what event it creates, which policy applied, and how that event can be trusted later.
Peer-to-peer events
Systems can exchange transaction events directly while avoiding another central data lake for every integration.
Field-level visibility
Only the fields needed for a task are exposed to the person, agent, workflow, or partner allowed to see them.
DPIA-ready audit
Events carry the context needed to know what happened, what was visible, and what policy applied at the time.
MCP and tools
Agent-facing docs, MCP server patterns, and implementation guidance help teams test TN-Proto inside real workflows.
From roadblocks to protocol
Common AI roadblocks pointed to the same evidence layer.
TN-Proto comes from our analysis of common AI adoption roadblocks: fragmented access, assumed trust, scattered permissions, and agent workflows that need reliable transaction context and reviewable evidence.
Firms want AI systems that can move across applications, reason over operational context, and trigger useful work. The weak point is often the event layer: permissions are scattered, logs are too blunt, source systems disagree, and agents create new questions about who saw what.
That pattern pushed us toward TN-Proto. It gives teams a repeatable way to test secure event flow, field-level visibility, and agent-readable transaction context across integration work.
Signals
Data movement is where value gets stuck.
Roadblock analysis often shows that teams know the use case before they have a trusted way to move data through the workflow.
Protocol
Events need identity, policy, and audit evidence.
Agentic systems need transaction events that carry enough context to be trusted, reviewed, and shared safely.
Tools
Teams need a way to try the pattern.
The hosted tools, docs, and agent skills make it easier to test the protocol in real integration work.
Agentic AI governance
Where TN-Proto fits.
The same questions show up in AI governance, MCP server design, data access governance, and AI workflow automation.
What does data governance need for agentic AI?
It needs event-level context. TN-Proto connects identity, policy, field-level visibility, data lineage, and audit evidence to the transaction event itself.
How does TN-Proto relate to MCP servers?
MCP servers give agents a way to use tools. TN-Proto gives those tool calls a secure event path with transaction context, access control, and reviewable evidence.
How can TN-Proto support DPIA work?
TN-Proto supports DPIA work by keeping evidence about purpose, event context, field-level visibility, access policy, and audit trails close to the transaction event.
Where does secure data sharing fit?
Secure data sharing is the operating layer for useful AI agents. TN-Proto helps teams test event-driven architecture patterns where agents, workflows, and partner systems can exchange only the data needed for the task.
DPIA-ready evidence
Designed for the questions privacy and governance teams ask.
Under GDPR, a Data Protection Impact Assessment is about understanding high-risk processing before it scales. Agentic AI makes that harder because tool calls, prompts, retrieved fields, and workflow events can move quickly across systems.
TN-Proto gives enterprise teams a cleaner evidence layer for those assessments. Each meaningful transaction can carry the purpose of the exchange, the actor or agent involved, the fields exposed, the policy applied, and the audit trail needed to review the event later.
Purpose
Why did this event happen?
Transaction context helps teams connect an agent action to a business purpose, workflow state, or approved operational need.
Visibility
Which data was exposed?
Field-level visibility supports data minimisation and helps separate useful access from broad system exposure.
Evidence
What can be reviewed later?
Policy, identity, event lineage, and audit trail data travel with the transaction, giving governance teams something concrete to inspect.
TN-Proto access
Start with the audit trail.
TN-Proto has a Free tier for exploring the protocol, a Business tier that is currently free, and a planned Enterprise tier for governed deployment work.