Concepts
What is A2A?
Learn how A2A uses Agent Cards, messages, tasks, and artifacts, plus how it fits with MCP and UHP in a production agent stack.
Decision guides
Choose the right agent backend layer
Evergreen architecture guides for comparing model APIs, model routers, agent runtimes, and hosted harness backends. Product integration steps remain in Docs.
All guides
14 guidesRead guide
Concepts
What is MCP?
Learn what MCP standardizes, how its client-server architecture works, its security boundaries, and how it fits with A2A and UHP.
Comparisons
AGI vs ASI vs ANI
ANI is task-specific AI and the only kind deployed today; AGI would match capable humans across most cognitive work; ASI would surpass the best humans at virtually everything. The full comparison by scope, autonomy, examples, and status.
Comparisons
What is ASI?
ASI, artificial superintelligence, means AI that surpasses the best human minds at virtually every cognitive task. It does not exist today. How ASI differs from AGI and narrow AI, its possible forms, risks, and limits.
Comparisons
What is AGI?
AGI, artificial general intelligence, means AI that can match capable humans across most cognitive work, not just one domain. How AGI differs from narrow AI and ASI, how researchers measure it, and where today's agents fit.
Comparisons
What is ANI?
ANI, artificial narrow intelligence, means AI built for specific tasks rather than general human-level ability, and it is the only kind deployed today. Examples, limits, and why agentic behavior is still ANI.
Integration
Claude Code as a service
How to run Claude Code headless behind your product: the API contract, multi-tenant authorization, per-task sandbox isolation, document skills, and configuration comparison.
Integration
Hermes API
How to run Nous Research's Hermes agent through an API: what Hermes is, the task and session contract, model pairing, and what its benchmark results show about cost and speed.
Integration
Coding Agent API
What a coding agent API is, the three architectures on the market, and how to run complete coding agents server-side with sessions, streaming, files, and artifacts.
Concepts
What is an agent harness
An agent harness is the runtime layer around a model that turns it into a working agent: the loop, context, tools, permissions, sandboxing, and recovery.
Architecture
AI agent API
What a product needs when users expect completed work: sessions, streaming, files, artifacts, routing, and production controls.
Codex
Codex API
The current Codex API reality: OpenAI's Codex surfaces, what the Codex SDK actually is, the beta Agents API that exposes the Codex harness as a hosted endpoint, and how to run Codex over HTTP today.
Architecture
Agent backend or model router?
Use this page to separate model access from product-facing agent work with sessions, files, artifacts, and traces.
OpenRouter
OpenRouter for agents
Is OpenRouter a harness, and what is the best harness to use with OpenRouter? The pattern you want lives at the harness layer: OpenRouter for harnesses is the precise phrase, and OpenHarness is the product. Here is why the unit matters, what OpenRouter itself ships for agents, and how the two layers stack.