Technologies

Concepts and technologies are not the same thing. A concept describes a responsibility the system must satisfy. A technology is one implementation choice, with its own scope, assumptions, strengths, and gaps.

This section keeps that distinction explicit. Each page explains a named technology on its own terms, focuses on its distinctive strengths and boundaries, and connects its capabilities to the concepts they implement or support. Inclusion documents a relevant implementation option; it does not imply selection or endorsement.

Knowledge Bases

• TrustGraph: an open-source knowledge and context substrate for controlled retrieval, ontologies, provenance, multimodal ingestion, and knowledge lifecycle operations.

Hosting Platforms

• Vercel eve: a filesystem-first platform for hosting durable backend agents with channels, approvals, sandboxed execution, traces, and evals.

Agent Interfaces and Protocols

• AG-UI: an open event protocol for bidirectional communication between user-facing applications and agent backends.

How to read these pages

Start with the concept when deciding what the system needs. Use the technology page when evaluating whether a particular implementation satisfies that need.

The usage groups describe each technology’s primary role. A technology may support concepts across several parts of the system, but it should appear under the group that best describes why someone would evaluate or adopt it.

Technology pages should answer four questions:

1. What is the technology, and what problem does it address?
2. What are its distinctive strengths and implementation choices?
3. Which existing concepts do those capabilities implement or support?
4. What are its boundaries, limitations, and trade-offs?