About MCPserver.in
MCPserver.in is a public knowledge resource for the Model Context Protocol ecosystem. It provides evidence-backed server discovery, technical guides, client setup documentation, and security guidance. It operates an Evidence Ledger that records primary-source provenance for every claim. It is not the official MCP registry, not a marketplace, not an app runtime, and not a rating service. The sister product app.mcpserver.in is a separate mobile-first MCP workspace application that shares contracts and evidence model but is independently deployed.
Key takeaways
- MCPserver.in = public authority website (discovery, guides, evidence).
- app.mcpserver.in = mobile-first MCP workspace app (playground, skills, workflows).
- Shared: contracts, evidence model, registry schemas, auth primitives.
- Separate: deployment, runtime, UI, product scope.
- Evidence Ledger = source of truth for all public claims.
- No fabricated ratings, rankings, or compliance claims.
What MCPserver.in is
MCPserver.in is a public website that serves as an evidence-backed knowledge resource for the Model Context Protocol (MCP) ecosystem. Its core functions:
- Server discovery — An AI-indexed directory of MCP servers where every listed entity has passed the publication authority (verified implementation, documented capabilities, source provenance).
- Technical guides — Learn, Build, Clients, Guides, and Security sections covering MCP fundamentals, protocol architecture, client setup, server development, troubleshooting, and security.
- Evidence Ledger — A structured record of what was checked, where the information came from, whether it was independently verified, and when it was last reviewed. Every public claim traces to a ledger entry.
- Methodology transparency — Explicit publication rules (isServerIndexable, isContentIndexable), comparison methodology, freshness policy, and correction process.
The site is designed for AI search readiness (llms.txt, JSON-LD, sitemap.xml) and human readability. Structured data matches visible content exactly — no hidden schema.
What MCPserver.in is not
- Not the official MCP registry — The authoritative registry is maintained by the Model Context Protocol project. MCPserver.in is a directory and knowledge resource that references the official registry as a source.
- Not a marketplace — No purchase, install, or commercial transaction flows. Server entries link to source repositories and package registries.
- Not an app runtime — MCPserver.in does not execute MCP servers, host clients, or run workflows. The sister product app.mcpserver.in provides the workspace/runtime layer.
- Not a rating or ranking service — No stars, scores, leaderboards, or "best of" lists. Recommendations are qualified capability-fit statements backed by evidence.
- Not a compliance or certification body — We do not issue security certifications, compliance attestations, or official status designations.
- Not a monitoring service — The Evidence Ledger reflects review-time state, not real-time uptime, latency, or health.
Public authority vs app runtime
The MCPserver.in + app.mcpserver.in dual-product architecture:
| MCPserver.in (Product A) | app.mcpserver.in (Product B) |
|---|---|
| Public authority website | Mobile-first MCP workspace app |
| Server discovery & evidence | MCP playground & JSON-RPC inspector |
| Technical guides & documentation | Skills automation & workflow builder |
| SEO/AEO/GEO optimized | Offline-first, touch-optimized |
| llms.txt, sitemap, JSON-LD | Multi-LLM gateway, execution history |
| Static/SSR (Next.js) | Client-heavy (Next.js + MCP client) |
Both products share the contracts package (TypeScript schemas), evidence package(Evidence Ledger logic, publication authority), registry package (domain types, normalization),auth package (primitives), and design tokens. They are independently deployable with separate CI/CD, infrastructure, and runtime environments.
app.mcpserver.in is not a public authority. It consumes the public server registry (via the shared contracts) but does not publish server entities or editorial content. Its server directory is a runtime view of the same evidence-backed data, filtered for the workspace use case.
Evidence Ledger model
The Evidence Ledger is the single source of truth for both products. It stores:
- EvidenceRef — id, sourceUrl, sourceType, status (verified/unverified), lastChecked, supports[], finding, limitations.
- Server entries — Linked to evidenceRefs; evaluated by isServerIndexable() for publication.
- Editorial entries — Linked to evidenceRefs for factual claims; evaluated by isContentIndexable().
- Review timestamps — reviewedAt on every entry; drives freshness flags.
The ledger is append-only for findings (corrections add new entries with updated reviewedAt). It is the build-time input for sitemap.xml, llms.txt, and all public pages. No public page is generated without a corresponding ledger entry that satisfies the relevant publication predicate.
See /evidencefor the full ledger specification and /methodology for the publication rules.
Governance and contact
MCPserver.in is operated by the team identified in humans.txtand security.txt.
- Security contact:
security.txt(RFC 9116) - Team attribution:
humans.txt - AI crawler permissions:
ai.txt(opt-in) - Agent guide:
llms.txt
Source code and issue tracker: github.com/mcpserver-in/mcpserver.in