Urgent.News

What's breaking now, across thousands of outlets.

Tech

Model Context Protocol Documentation: Spec, Schema, Changelog

Short answer: Model Context Protocol documentation is versioned by date, not by number. Each revision has its own specification pages, a schema published beside them, and a changelog recording what moved since the previous revision. Read the revision your client negotiated, use the schema when a field's shape matters, and read the changelog before you upgrade anything. Key takeaways Three…

Model Context Protocol documentation consists of three essential documents: a specification, a schema, and a changelog. The specification outlines the required elements, while the schema defines the shape of each field, and the changelog records changes and their corresponding revisions. Revisions are identified by dates, and each revision has its own specification, schema, and changelog.

A citation without a date refers to the site, not a specific document. Older revisions remain accessible. Clients that negotiate a specific revision are not affected by newer revisions unless the server stops adhering to the older text. The schema is preferable to prose for precise information about fields, as it provides exact enum values, optional fields, and error codes.

The changelog is the sole reliable source for upgrading, as it explains what changes were made and the reasoning behind them. The Model Context Protocol's documentation set comprises four components: the specification, schema, changelog, and registry. The specification, published per revision under modelcontextprotocol.io/specification/revision/, contains chapters for lifecycle, transports, authorization, and server/client features.

The schema is the TypeScript source of wire types, published alongside each revision. The changelog lists changes since the previous revision, along with proposal identifiers. The registry is a separate metadata service for discovering servers, distinct from the other documentation surfaces. When disagreements arise, follow this order: specification for requirements, schema for shapes, changelog for history, and registry only for discovery.

A professional reading order for the specification includes five pages: the overview, versioning, lifecycle, transports, and the primitive used. A professional reading order for the specification is five pages long, in this order: the overview for the protocol, versioning for agreement on revisions, lifecycle for connection behavior, transports for message travel, and the primitive page.

When reading vendor documentation, such as Databricks' gateway docs, distinguish between vendor implementation and protocol requirements. Vendor documentation should describe what the vendor configures and how clients must send requests, while the specification defines the protocol requirements.

Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

Read the original at dev.to →

More in Tech

RAG Architecture Diagram: the Boxes and What Each Arrow Costs

Short answer: A RAG architecture diagram is five boxes and the arrows that join them - an index you build, a retriever that turns a query into candidates, a reranker that orders them, a context…

  • RAG architecture consists of five boxes and six arrows
  • Loop controller incurs most of the costs in the process
  • Context builder's effect measurable per single request

More from Wednesday 30 September →