Submission: Unyly
URL: https://unyly.org
Type: Catalog / discovery tool (not an MCP server itself — a registry that indexes MCP servers across the ecosystem)
License: Catalog content public, code partially open-sourced
What it does
Unyly indexes 15,000+ MCP servers from:
- Official MCP Registry (this one)
- Smithery
- PulseMCP
- Glama
- npm (
@modelcontextprotocol/* and related)
- Community submissions
Catalog data is normalised across sources: install commands, environment variables, permission scope, repo health, npm publish date, security considerations. Each server gets a unique URL with structured metadata (Schema.org SoftwareApplication, BreadcrumbList) — improves discoverability via Google + AI search engines.
Install flow
We solved the "config-file friction" problem. Users click "Install" → we detect their client (Claude Desktop / Claude Code / Cursor / VS Code Cline) and inject the right config block. No manual JSON editing.
Why this matters for the Official Registry
- We pull from the Official Registry (your source-of-truth feed is read in our daily ingest job)
- We send users back — every listing has a "View on Official Registry" link with the canonical registry URL.
- We add a discovery + install UX layer that complements the registry's role as canonical source of truth.
Happy to add additional fields if the Official Registry exposes them. We currently parse: name, slug, repo, npm package, install command, env vars, description, tags, license.
Tech
- 15,000+ MCPs indexed (~6,000 unique after dedup)
- Daily ingest pipeline (Go workers)
- SQLite catalog + Drizzle ORM + Next.js 16 frontend
- Bilingual EN/RU
- Free tier covers all browse + install operations
Contact
GitHub: @FasadSalatov
Email: salatiksama@gmail.com
Site: https://unyly.org
Submission: Unyly
URL: https://unyly.org
Type: Catalog / discovery tool (not an MCP server itself — a registry that indexes MCP servers across the ecosystem)
License: Catalog content public, code partially open-sourced
What it does
Unyly indexes 15,000+ MCP servers from:
@modelcontextprotocol/*and related)Catalog data is normalised across sources: install commands, environment variables, permission scope, repo health, npm publish date, security considerations. Each server gets a unique URL with structured metadata (Schema.org SoftwareApplication, BreadcrumbList) — improves discoverability via Google + AI search engines.
Install flow
We solved the "config-file friction" problem. Users click "Install" → we detect their client (Claude Desktop / Claude Code / Cursor / VS Code Cline) and inject the right config block. No manual JSON editing.
Why this matters for the Official Registry
Happy to add additional fields if the Official Registry exposes them. We currently parse: name, slug, repo, npm package, install command, env vars, description, tags, license.
Tech
Contact
GitHub: @FasadSalatov
Email: salatiksama@gmail.com
Site: https://unyly.org