Skip to content
PointSav Digital Systems

Last updated · 9e782ef0

MCP as substrate protocol

MCP-as-Substrate-Protocol designates the Model Context Protocol (MCP) as the wire contract for all inter-service composition in the platform. Every Ring 1 and Ring 2 service exposes an MCP server interface as its primary external contract. The Doorman (service-slm) is the MCP gateway. Customer extensions plug in as additional MCP servers. This is not a preference or an implementation convenience — it is a structural decision.

Why MCP is substrate-level

The Model Context Protocol has become the industry standard for AI-native application composition. It defines a stable, machine-readable interface between clients, servers, and host processes. 1 The platform adopts it at the substrate level because the alternative — bespoke per-service wire formats — accumulates versioning debt, per-pair contract testing costs, and custom client implementations in every consumer. MCP is the coordination point that prevents that drift.

The practical result: a customer-built agent, an IDE extension, and the platform operator TUI all interact with the same service interfaces using the same protocol. There is no "developer API" distinct from the "user API." The wire contract is unified.

How the three-ring architecture maps to MCP roles

MCP defines three roles. The platform's architecture maps onto them directly:

MCP Server — each Ring 1 and Ring 2 service: the file system ledger, the knowledge runtime, the extraction service, the input service, the people service, the email service, and the egress service. When Phase 5 delivers the marketplace and settlement services, those are MCP servers too.

MCP Client — the Doorman (consuming Ring 1 and Ring 2 services as tools during inference); the operator TUI; customer-built agents; IDE extensions.

MCP Host — the Doorman for inference-traffic flows; the operator TUI for operator-traffic flows.

The Doorman is both MCP Client (calling service-content for graph grounding) and MCP Host (presenting the unified inference interface to upstream callers). The reuse is deliberate: the same process that holds inference credentials also mediates tool composition.

Tool semantics

Each service exposes a small set of named MCP tools. service-content exposes graph query, graph mutation, vector search, and temporal query. The file ledger service exposes ledger append, ledger query, and checkpoint. The extraction service exposes entity extraction and classification. The people service exposes person lookup, upsert, and relationship query.

The marketplace service (planned Phase 5) is intended to expose listing creation, listing query, and transaction initiation as MCP tools. Customers extending the platform add new MCP servers that expose their own vertical-specific tools. The Doorman discovers new tools at session start via MCP's describe mechanism; no core code changes are required to accommodate a new extension.

The Doorman as MCP gateway

When any caller — task session, operator TUI, or customer agent — sends an inference request to the Doorman, the Doorman is simultaneously a client (calling service-content for graph context per knowledge-graph-grounded-apprenticeship) and a gateway (routing the request to the appropriate compute tier). This mediation point is where graph grounding, audit logging, tier selection, and compute-boundary enforcement all converge.

Customer extensions

A customer with a vertical-specific data source writes an MCP server that exposes that system's data as tools. The Doorman discovers and uses those tools in subsequent inference calls alongside the built-in tools. This is the structural opening for the vertical-seed-packs-marketplace: each planned pack is intended to ship with a reference MCP server for that vertical's typical data sources.

Composition with Code-for-Machines First

MCP is the structural realization of the code-for-machines-first principle. Every service contract is machine-readable and introspectable at runtime. Human-facing surfaces — the TUI, web interfaces — are clients of the same MCP interfaces that any other agent would use. There is no human-only data surface in the platform.

OpenAI-compatible HTTP boundary preserved

The Doorman's existing OpenAI-compatible HTTP interface is preserved alongside MCP. Third-party clients that use the OpenAI SDK continue to work without modification. MCP is added as the substrate-level composition standard, not as a replacement for existing compatibility.

See also

  1. Anthropic. 'Model Context Protocol Specification.' modelcontextprotocol.io, 2025. https://modelcontextprotocol.io/specification/2025-11-25 ↩

Important Information

Important Information

Corporate structure. PointSav Digital Systems ("PointSav") is a trade name of Woodfine Capital Projects Inc. ("Woodfine"). PointSav does not itself offer, sell, or solicit any security. Any securities offering associated with Woodfine's real-property direct-hold solutions is made exclusively by Woodfine, and only by means of the applicable Private Placement Memorandum.

No investment advice. This wiki's content is provided for engineering, operational, research, and development purposes. Nothing on this wiki constitutes investment advice or a solicitation to invest in any Woodfine partnership or direct-hold solution.

Intellectual property. The PointSav name, trade name, wordmark, and marks, together with all current and future PointSav- and Totebox-branded products, services, and offerings — and the software, source code, documentation, design system, and all related materials — are proprietary to Woodfine and its affiliates, except for components identified as open source. No rights are granted except as expressly set out in a written license or agreement. See TRADEMARK.md in this repository for the full trademark notice.

Open source components. Portions of the platform are made available under permissive open-source licenses identified in the accompanying repository. Use of those components is governed by their respective license terms.

No warranty; informational use. Content on this wiki is provided for general informational purposes only and does not constitute a representation, warranty, or commitment with respect to product functionality, availability, pricing, or roadmap. Some articles describe planned or intended features, capabilities, and milestones — language such as "planned," "intended," "targeted," "may," and "expected" marks this forward-looking content, which is subject to change and does not constitute a commitment regarding future performance.

Confidentiality. Where an article describes an operational or deployment detail that is not intended for public disclosure, that article is not published on this wiki. Content here is general-purpose engineering documentation, not customer-specific configuration.

Jurisdiction. Woodfine Capital Projects Inc. is organized in British Columbia, Canada. References to the Sovereign Data Foundation on this wiki describe a planned or intended initiative only, not a current equity holder or active governance body.

Changes to this notice. PointSav may update this notice from time to time; the version posted on this page governs.

Not a filing system. This wiki is not a securities filing system, an electronic disclosure repository, or a substitute for SEDAR+ or any other regulatory filing system. Formal securities filings are made through the applicable regulatory filing system, not through this wiki.

Full disclaimer. This notice supplements, and does not replace, the full Disclaimers article. In the event of any conflict, the full Disclaimers article governs.

Read the full disclaimer →