A Cascade alternative for teams who steer with the X-Matrix itself

Cascade's strength is that hoshin is one of several supported frameworks. That is also the trade: as of our July 2026 review it has no dedicated rotatable X-Matrix, and hoshin is not the object the product is organised around.

Everything we build assumes it is. The four quadrants, the constraint that correlations only exist between adjacent quadrants — so a vision reaches a KPI through an annual objective and an initiative, never directly — and a completeness score that judges the result out of 100 across five metrics: coverage, correlation density, orphan elements, complete vision-to-KPI chains, and the leading-versus-lagging indicator balance.

That specialisation is a real limitation if your organization runs OKRs in one division and scorecards in another. There is no honest way to tell you a single-framework tool serves that better.

Tapestry is genuinely good, and it is an assistant inside Cascade: you go to their application and it helps you there. As of our July 2026 review, we found no public REST API documentation and no MCP capability, so the AI is a feature of the product rather than a surface for yours.

We inverted that. The MCP server is the primary interface — MIT-licensed, published on npm and in the official Model Context Protocol registry, working on the free tier. The consequence is not a better chatbot; it is that the strategy becomes available in the tools where work already happens. Your agent reads the matrix in the middle of a conversation about hiring, notices the initiative it touches has no KPI, and says so. Nobody opened a strategy application to make that happen.

Since July 2026 the agent can also build: create_matrix, upsert_element and set_correlation mean an assistant turns a strategy document into a scored matrix in one conversation, instead of a human retyping it into a form.

• Cascade: Tapestry AI in-app; no public REST API docs or MCP found (July 2026); numerous native integrations (Tableau, Power BI, Jira, SAP).
• hoshin.app: MCP-first, 14 tools and 9 guided prompts, OAuth 2.1 with dynamic client registration for claude.ai and Cursor, per-user rights and full audit trail.

Cascade's compliance posture is stronger than ours on paper — SOC 2 Type II is a real audit and we do not have it. On data residency it offers multiple regions including the EU, which answers most enterprise requirements.

It does not answer the requirement that ends some deals outright: strategy that never sits on a vendor's servers at all. hoshin.app's sovereign local mode runs the same 14 tools and the same score against plain JSON files on your own machine — no account, no network, no telemetry, and MIT-licensed code you can audit before running. For consulting firms, law firms and anyone whose business is confidential method, that is not a preference; it is the precondition. Multi-region hosting is a different answer to a different question.

Where Cascade wins: multi-framework breadth, SOC 2 Type II, portfolio management, the maturity of a larger product, and an in-app AI that is more polished than anything we render in our own interface.

Where we win: the X-Matrix as the product rather than a template, a completeness score that returns a verdict instead of a dashboard, the agent as the interface with an open and registered MCP server, and a local mode where your data never leaves your machine.

Where we simply are not comparable: we are small. No SOC 2, no portfolio module, no customer-success organisation. If those decide your purchase, decide it in their favour.