Connection diagrams
How agents reach each connected system, and through which interface.
Every engagement ships with connection diagrams, endpoint specifications, validation rules and full flow documents. This is what that documentation covers.
Detailed technical documentation is produced as part of delivery and handed over with the build. Contact us for access to the current set for your environment.
How agents reach each connected system, and through which interface.
The REST, OData and Service Layer endpoints each integration uses.
The checks applied before any transaction is committed to a system of record.
End-to-end process flows for every automated workflow in scope.
Isolation model, authorisation rules and deployment gateways.
What is written to the audit trail, where it lives and how to query it.
Side panel UI with chat interface, live state updates and conversation history; an agent launcher for new tasks, sessions and templates; an options page for model settings and configuration.
A background service worker acts as the central orchestrator, handling message routing, task management, tab management, event handling and connection management.
The core multi-agent system. An executor orchestrates agents and manages the task lifecycle; a planner agent understands the task, creates the plan, self-corrects and breaks down steps; a navigator agent executes it against live pages.
Multi-model support across Anthropic Claude, OpenAI, Google Gemini, Ollama local models and any OpenAI-compatible endpoint.
Tabs, DOM, cookies, storage and navigation, the browser capabilities the agent uses to act on a page the way a person would.
SAP Fiori, ERP systems, websites, internal portals and any other URL-based application your teams already work in.
Cross-cutting services: storage and settings sync, action caching, LLM observability and tracing, usage analytics, resource monitoring and RPA integration for hybrid automation.
Tell us what your architecture or security team needs and we will send the relevant documentation.