If you want to work with the Meta-Primer agent directly, the easiest route is through the OpenAI Library on the web. All you need is a standard OpenAI account. The reason we point people there is simple: only the browser-based interface supports the full interactive workflow the agent relies on — long-form reasoning, multi-step sessions, and the richer UI components that keep the analysis coherent. API access and multi-client apps are great for lightweight or automated tasks, but they can’t yet deliver the depth, structure, or continuity required for the full experience.
When you access the agent through the OpenAI Library, you’ll also need a session key. This key connects your workspace to the correct configuration and keeps your interactions aligned with the version of the agent you’re meant to use.
Keys are unique and they are time limited, provided by the Project.
It’s a simple one-time step: enter the key when prompted, and the system locks everything in. Without it, the agent will load in its default mode, which won’t include the specialised tools, behaviours, or settings you expect.
👋 Welcome to XPlain
Your meta-reasoning framework for structured AI reasoning
What is XPlain?
XPlain is a meta-primer framework that provides explicit reasoning architectures for AI systems. Think of it as giving AI a structured thinking process – making reasoning transparent, consistent, and auditable across different platforms and use cases.
Instead of hoping AI “figures it out,” XPlain provides clear cognitive frameworks that guide reasoning from problem understanding through to validated outputs.
⚠️ Important: Session Code Required
Before we begin any work, you’ll need to provide your session code. This ensures continuity across our conversation and maintains access to your specific XPlain configuration and history.
If you don’t have a session code yet, just ask – I’ll help you set one up!
What Can You Do Here?
This is your XPlain workspace. You can:
- Explore the framework – Learn about meta-primers, topical primers, and how XPlain structures reasoning
- Create primers – Build reasoning frameworks for specific domains (GRC, risk assessment, compliance, etc.)
- Test and validate – Run reasoning tests across multiple AI platforms to verify consistency
- Review results – Analyze Romers (reasoning outputs) and assess quality
- Iterate and improve – Refine primers based on empirical results
- Get guidance – Ask questions about XPlain methodology, best practices, and implementation
🚀 Ready to Start?
Here are some ways to begin:
- “Explain XPlain architecture” – Learn about the Meta-Primer structure
- “Show me a primer example” – See what a topical primer looks like
- “I want to create a primer for [domain]” – Start building your own
- “What’s the difference between XR and XA?” – Understand schema variants
- “Help me test reasoning consistency” – Set up validation testing
Key Concepts to Know
Meta-Primer (XR): The universal core reasoning structure – theory-neutral and domain-agnostic
Topical Primers: Domain-specific implementations of the meta-primer for specialized use cases
Romers: Reasoning outputs – the structured results generated when AI uses XPlain frameworks
XPlain-A (XA): AAMAS-aligned interpretation layer that maps XPlain to agent theory terminology
Philosophy
XPlain is built on the principle that explicit reasoning architectures make AI systems more reliable, transparent, and accountable. Rather than treating AI as a black box, XPlain provides visible cognitive scaffolding that:
- Makes decision processes auditable
- Achieves 90-95% consistency across AI platforms
- Enables meaningful validation and verification
- Supports collaborative human-AI reasoning