The concern
If TT charges OpenAI / Anthropic / Google significant AI-vendor revenue (£5M+/year plausible) and the MCP infrastructure is Sourcemeta-built, Juan can reasonably argue the 50/50 spirit of the existing agreement (§5 hosted-validation revenue) should extend to the AI-vendor deal. Strictly that’s not what §5 says — but it’s a conversation, not a certainty.
The existing agreement (from memory)
Signed BoldSign, effective Tuesday 14 April 2026:
- §2 — Sourcemeta is “Validation Technology Partner” (narrow scope)
- §3 — exclusivity on validation tech
- §5 — 50/50 on hosted validation service revenue
- §9 — Apache 2.0 schemas
- §14 — 30-day exit
MCP auto-generation is outside §2 scope. §5 is specifically on hosted validation service, not MCP or AI-vendor deals. Best-read conclusion: the existing agreement does not give Juan revenue share on MCP or AI-vendor deals. But commercial scope disputes are common in vendor relationships and warrant explicit clarification.
The two-tier MCP architecture (strategic answer)
Tier A — Apache INHERIT MCP via Sourcemeta One auto-MCP:
- Exposes schemas, reference data (rule values, statute references)
- Content is Apache 2.0 — commercially low-value
- Free to consumers — adoption driver
- Sourcemeta-delivered, no dispute because no revenue
Tier B — InheritKit MCP (TT-built, TT-hosted):
- Exposes AGPL+dual-licensed InheritKit operations (IHT math, faraid distribution, intestacy walks, cross-layer consistency, workflow orchestration)
- Commercial revenue lives here (Route 1 embedded-licence + Route 2 metered API)
- TT keeps 100% of revenue — InheritKit is outside Sourcemeta agreement scope
Why this is defensible to Juan
- Existing agreement literally scopes Sourcemeta to validation, not InheritKit
- InheritKit is a separate product category Juan already agreed to at signing (InheritKit design spec §5 analysis confirmed no conflict)
- Sourcemeta still benefits — Tier A drives INHERIT adoption; INHERIT adoption drives InheritKit demand; InheritKit demand drives hosted-validation demand. Partnership is still commercially valuable.
- Clean commercial boundaries are standard professional practice; clarifying them is healthy, not adversarial.
Juan-conversation framing
Recommended approach:
- Internal TT analysis first (via Phase 1 subject 37 in v3.6 handoff prompt)
- Then structured conversation with Juan BEFORE v4.0.0 ships, while leverage exists
- Frame: “I want to clarify scope on AI-vendor revenue so we both know where we stand.” Not adversarial.
- The existing agreement is favourable to TT on this point; the conversation confirms what already is true, not negotiates TT’s position from weakness.
- Keep the relationship healthy — Juan has delivered materially (Blaze, toolchain, doc generator, canonicaliser, auto-MCP plan). Goal is commercial clarity, not relationship damage.
Alternative structures if two-tier is unworkable
- Amended supplemental agreement specifically about MCP / agent-readiness revenue
- Joint-venture structure for AI-vendor sales with agreed split
- Negotiated 20/80 or 10/90 split on Tier A-derived revenue only (if Tier A ever generates revenue)
Each has costs/benefits; the research session evaluates.
Action items
- Internal analysis (Phase 1 subject 37 of v3.6 handoff prompt) — literal agreement reading, two-tier architectural feasibility, alternative structures
- Engineering decision — confirm two-tier MCP architecture feasible; TT capacity to build Tier B
- Juan conversation — structured, collegial, pre-v4.0.0-launch, clarifies scope without negotiating from weakness
- Supplemental agreement (if needed) — document whatever clarification emerges in writing, append to existing BoldSign agreement
Why pre-v4.0.0 matters
Until v4.0.0 ships publicly:
- Leverage is with TT (no external users yet)
- Juan has delivered value but not yet captured AI-vendor revenue
- Scope clarification is easy
After v4.0.0 ships:
- External users create facts on the ground
- If AI-vendor deals start flowing before scope is clarified, the conversation is harder
- Delayed scope clarification looks like bad-faith ex-post negotiation
Sensitivity note
Juan is an active, valued partner. The goal of this analysis is commercial clarity, not relationship damage. A clean two-tier architecture that preserves Juan’s upside on validation-adjacent work while protecting TT’s AI-vendor revenue is the correct frame. Any recommendation or conversation that frames Juan as adversarial is wrong.