Running AI agents inside WSL?
Tell us about your WSL setup. We’ll map the supported traffic paths and configuration needed for visibility and control.
THE VERAX PLATFORM
Every AI request, checked against your organization’s policies.
YOUR ORGANIZATION
Identity, AI and tools. Connected by your policies.
IDENTITY CONTEXT
Finance analyst
Authenticated identity
Connected identity providers
VERAX POLICY ENGINE
Vera validates the next action before it reaches a tool.
AI REQUEST

YOUR ORGANIZATION’S POLICY
Checks every request
Authenticated identity
The identity context travels with the AI request.
Identity is connected to the requested action and resource.
01 SELF-HOSTED. UNDER YOUR CONTROL.
Run inspection and policy enforcement inside infrastructure you control. Your team defines what information and actions are permitted before approved requests continue to authorized AI services.
REMOTE REQUEST
Outside your organization.
INTERNAL REQUEST
Inside your environment.
ROUTE
Bring supported traffic into the policy path.
CHECK
Allow, redact or block under your rules.
APPROVED CONTENT
AI apps · embedded AI
Code assistants · platforms
Follow a request through your environment.
02 CONTROL THE ACTION
Let AI summarize a report without giving it permission to rewrite the numbers. Verax checks each supported action.
AI ASSISTANT
VERAX POLICY ENGINE


Finance / Reports
SOURCE DOCUMENT
Q3 · Finance team
Category
Amount
Travel
$24,800
Marketing
$18,600
Operations
$42,900
Quarter total
$86,300
Original report stays unchanged
Vera checks the request before the assistant accesses the report.
03 VERAX IN YOUR ORGANIZATION
Connect AI work to the person, information and system behind it. Apply your rules locally before approved context or actions continue.
Choose a task, then change the example policy to compare what happens.
Review a repository
GitHub · READ
01
Summarize project notes
Notion · READ
02
Update a backup schedule
AWS · WRITE
03
ACCOUNTABLE IDENTITY
Operations analyst

EXAMPLE POLICY
SELF-HOSTED VERAX
Your infrastructure
Summarize these project notes.


VERAX POLICY ENGINE
Remove restricted fields before AI receives the context.
Request received
CONNECTED SYSTEM
READ
checkout-service
SOURCE REPOSITORY
export function checkout() {
return createOrder();
}
Restricted config
demo-only-token
Review code without changing the repository.
Launch preparation
SOURCE DOCUMENT
Status: in review
Next step: finalize the launch checklist.
Internal contact
team@company.example
Access token
demo-only-token
The saved document stays unchanged.
Backup schedule
SOURCE SETTING
Current schedule
Daily
Requested change
Daily → Weekly
The setting changes only if the example policy allows WRITE.
WHAT CONTINUES TO AI
Verax evaluates the request before it continues.
Repository review ready.
Restricted config
[REDACTED]
Project context is ready for the AI summary.
Internal contact
[REDACTED]
Access token
[REDACTED]
No change made.
Schedule after policy
Daily
ATTRIBUTED DECISION
Identity + request + resource
ACTION
READ
RESOURCE
Project notes
POLICY RESULT
Checking policy
The robot brings the request to Vera for a policy check.
01
Set the information and actions each supported AI path can use.
02
Evaluate identity, content and operation in infrastructure you control.
03
Allow, remove restricted context or block—then review the decision.
04 ATTRIBUTION & AUDIT
Connect supported activity to its responsible identity. Review role permissions, authentication context, prompt context and the policy decision.
REQUESTED BY
Finance analyst
Attempted action: update report
ATTRIBUTED DECISION
The attempted change is denied. The source report stays unchanged.
ILLUSTRATIVE AUDIT RECORD
AUTHENTICATION
Authenticated via corporate IdP
ROLE PERMISSIONS
Read only
PROMPT CONTENT
“Summarize quarterly expenses.”
Illustrative record. Available evidence depends on supported integrations and recorded context.
Identity and context stay connected to the action.
COVERAGE & DEPLOYMENT
Tell us about your WSL setup. We’ll map the supported traffic paths and configuration needed for visibility and control.
Support for internal or self-hosted models is not available yet. Tell us about your deployment requirements.