The library,
inside your assistant
Search, browse and unlock prompts from Claude or any MCP client. A prompt unlocked there is unlocked here — one entitlement, both surfaces, and the same free-unlock allowance.
- Transport
- Streamable HTTP
- Auth
- OAuth 2.1 + PKCE
Point your client at the endpoint below. Dynamic client registration is supported, so there is nothing to create in a dashboard first — the client registers itself, sends you here to approve it, and receives a token scoped to this resource.
https://webbzer-frontend.vercel.app/api/mcp
{
"mcpServers": {
"webbzer": {
"type": "http",
"url": "https://webbzer-frontend.vercel.app/api/mcp"
}
}
}search_prompts
Full-text search over the whole catalogue.
list_prompts
Browse with the same filters the site uses.
get_prompt
Unlock and return a prompt — the same gate as the web unlock.
related_prompts
Neighbours of a prompt you already have.
list_categories
The category taxonomy.
account_status
Plan, limits, and what you have left.
Denials come back as structured results carrying an upgrade CTA, not as protocol errors — your assistant can explain the paywall instead of reporting a failure.
Read live from the server. If a client cannot find these, it is pointed at the wrong origin.
How authorization works
- 01Your client discovers the authorization server from the .well-known documents above.
- 02It registers itself dynamically and starts an authorization request with a PKCE challenge.
- 03Your browser opens our consent screen. You sign in and approve the client by name.
- 04The client exchanges its single-use code for a token bound to this resource.
- 05Every tool call runs against your plan — the same limits, counters and entitlements as the site.