Guide
Five controls to establish before enabling AI coding agents
An introduction to authorised project access, isolated execution, policy, human review and evidence.
BatonX / Resources
Practical guidance for governed AI coding.
Guides, explainers, checklists and architecture briefs for introducing coding agents with clear project boundaries, human review and customer-owned control.
Control. Context. Evidence.
The launch collection.
Read the five-control guide, model-gateway explainer, pilot preparation checklist, shared-governance architecture brief and security note.
Explainer
Why model gateways alone cannot govern local agent execution
Understand the distinction between governing model requests and governing the wider coding-agent session.
Checklist
Preparing for a customer-managed BatonX pilot
Prepare a representative project, an approved model endpoint, the developer workflow and operational responsibilities.
Architecture brief
CLI and read-only Web Chat under one governance layer
Explore two experiences with shared project entitlements, policy, approved model access and audit foundations.
Security note
What does BatonX record—and what does it leave out?
Understand the difference between retained security metadata and source code, prompts and responses that are not centrally retained by default.
Explore the platform today.
The product pages below are available now. They provide public, high-level explanations while the resource collection is prepared.
Why BatonX
Understand why the governance scope extends beyond a model request.
Explore the control scopeCustomer-managed deployment
Understand the customer environment and approved model-destination boundary.
Explore deploymentOne platform, two experiences
Explore Developer CLI and read-only Web Chat under shared governance.
Explore the platformSecurity and evidence
Review public security principles, data minimisation and human accountability.
Explore securityLet’s discuss the technical details
Contact us to explore how BatonX could fit your development environment. We can discuss agent compatibility, deployment needs and your security requirements.
BatonX
Useful public guidance. A clear disclosure boundary.
Public resources explain risks, outcomes and high-level controls. Detailed enforcement design and customer-specific operational material belong in an appropriately confidential, qualified review.
Explore the security-review approachApply the conversation to your own environment.
Discuss a pilot around your project boundary, supported agent, approved models and review workflow.
Discuss a pilot