Access
Bind every request to the real human, agent, client, tenant, and source-system permission ceiling.
Explore accessThe governed runtime for enterprise agents
Let agents act.
Keep control.
Onterix is building the customer-controlled execution and data-protection layer for enterprise AI—preserving real permissions, inspecting sensitive data, approving exact actions, and recording evidence across every agent interface.
GOVERNED TRANSACTION
Interactive trace“Prepare the Acme renewal brief, update the forecast, and send the customer follow-up.”
STEP 01
Bind the real actor and client
Maya asks Codex to prepare a renewal brief and update Salesforce. Onterix binds her identity, tenant, client, and exact capability versions before any source is touched.
Identity established
- Actor
- Maya Chen · Revenue Ops
- Client
- Codex · verified
- Capabilities
- 4 requested · 3 visible
STEP 02
Inspect the data boundary
Source labels and configured protection providers classify the assembled context. Customer PII is tokenized and restricted pricing is kept on the approved model route.
Permitted derivative created
- Sources
- Drive · Slack · Salesforce
- Findings
- 2 PII · 1 confidential
- Decision
- Transform + force route
STEP 03
Approve the exact mutation
The proposed CRM change and customer email are separate transactions. Policy allows the forecast update and requires a revenue-owner approval for the external message.
Payload sealed for review
- CRM diff
- Stage 2 → Stage 3
- 1 recipient · 1 attachment
- Approval
- Revenue owner · 30 min
STEP 04
Revalidate, execute, and prove it
Immediately before execution, Onterix rechecks policy, permissions, source revisions, and the approved payload. The result is idempotent and the full decision chain becomes evidence.
Executed with evidence
- Execution
- 2 actions · idempotent
- Policy
- Revalidated at 14:32:08
- Trace
- otr_01JQ9F…
One governed path for the tools your teams already choose
ClaudeChatGPTCodexCursorInternal agentsThe trust boundary
Control the full transaction.
Most platforms solve one part of enterprise agency. Onterix is designed to connect identity, data protection, action governance, and evidence at the moment an agent does real work.
Data
Inspect, classify, minimize, route, transform, or quarantine sensitive content at the boundaries you control.
Explore dataAction
Turn generated intent into typed, policy-checked transactions with exact-payload approval when risk demands it.
Explore actionEvidence
Connect identity, sources, transformations, decisions, approvals, and execution in one reconstructable trace.
Explore evidenceProtocol-neutral by design
MCP is the entry point. Governance is the product.
MCP, CLI, API, SDK, workflows, and future agent protocols should all enter the same capability runtime—not create separate policy engines, credentials, or audit stories.
See the architectureClaudeChatGPTCodexCLIInternal agents
ONTERIX RUNTIMEOne governed execution path
IdentityData policyApprovalEvidence
Company systemsKnowledgeModelsWorkflows
Same-or-stricter access
The agent never becomes a permission shortcut.
Onterix is built around a simple invariant: access through an agent must remain the same or stricter than access in the source system. Everything else follows from that rule.
Know who—and what—is asking
Authenticate the actor, delegated subject, agent, client, tenant, and runtime before discovery or execution.
Keep source permissions as the ceiling
An agent can narrow a user’s access. It can never silently expand it or substitute a broad service account.
Understand the destination
Source access does not imply permission to send content to every model, recipient, artifact, or external system.
Govern the exact action
Policy can allow, constrain, transform, route, require approval, quarantine, or deny a typed capability request.
Leave evidence, not a mystery
Every decision and mutation is tied to immutable versions, content hashes, source references, and a trace.
The Acme renewal
One question. Every control visible.
Our reference experience uses a synthetic company environment to make the product’s hardest promises testable: permission-aware retrieval, destination-aware protection, exact-action approval, and evidence.
GOVERNED TRANSACTION
Interactive trace“Prepare the Acme renewal brief, update the forecast, and send the customer follow-up.”
STEP 01
Bind the real actor and client
Maya asks Codex to prepare a renewal brief and update Salesforce. Onterix binds her identity, tenant, client, and exact capability versions before any source is touched.
Identity established
- Actor
- Maya Chen · Revenue Ops
- Client
- Codex · verified
- Capabilities
- 4 requested · 3 visible
STEP 02
Inspect the data boundary
Source labels and configured protection providers classify the assembled context. Customer PII is tokenized and restricted pricing is kept on the approved model route.
Permitted derivative created
- Sources
- Drive · Slack · Salesforce
- Findings
- 2 PII · 1 confidential
- Decision
- Transform + force route
STEP 03
Approve the exact mutation
The proposed CRM change and customer email are separate transactions. Policy allows the forecast update and requires a revenue-owner approval for the external message.
Payload sealed for review
- CRM diff
- Stage 2 → Stage 3
- 1 recipient · 1 attachment
- Approval
- Revenue owner · 30 min
STEP 04
Revalidate, execute, and prove it
Immediately before execution, Onterix rechecks policy, permissions, source revisions, and the approved payload. The result is idempotent and the full decision chain becomes evidence.
Executed with evidence
- Execution
- 2 actions · idempotent
- Policy
- Revalidated at 14:32:08
- Trace
- otr_01JQ9F…
NOTEThe Acme environment is a simulated product proof for the private preview. Connector and client readiness will be labeled as it advances from fixture to sandbox to verified.
Customer-controlled deployment
Bring the control plane. Keep the sensitive plane where it belongs.
The same runtime boundary is designed to support managed, dedicated, customer-data-plane, and fully self-hosted deployments without changing how agents discover and call governed capabilities.
| Planned mode | Control plane | Sensitive execution | Best fit |
|---|---|---|---|
| Managed shared | Onterix | Onterix regional cell | Fast evaluation and standard workloads |
| Managed dedicated | Onterix | Dedicated Onterix cell | Isolation, regional, and scale requirements |
| Customer data plane | Onterix | Customer environment | Customer-local credentials, content, and execution |
| Fully self-hosted | Customer | Customer environment | Disconnected or maximum-control environments |
“The neutral trust boundary underneath every enterprise agent.”
Build with us
Make governed agency real.
We are selecting design partners with serious agent workflows, sensitive systems, and a need for customer-controlled execution.