The full upstream README, mirrored here for reference. Install config, tool schemas, adoption signals, and an original overview live on the Snipzr listing page.
The official Snipzr MCP server. One hosted endpoint gives any MCP client thirteen tools to create, manage and measure short links: branded domains, batch creation, cookie-less analytics, campaign reports, UTM building and QR codes.
There is nothing to install or run. The server is remote (Streamable HTTP), stateless and available on every paid Snipzr plan. This repository is its public home: setup, the tool reference and the registry metadata. The server itself is part of the Snipzr platform, and its tools are thin adapters over the same handlers that serve the public REST API, so the two surfaces always behave identically.
1. Connect with sign-in (recommended). Most clients need no token at all: point them at the server address and they open a Snipzr sign-in. The consent screen shows exactly which permissions the application gets and which of your workspaces it may act in; untick anything you do not want, and disconnect it later from Settings.
Claude Code, two commands:
Claude on the web, desktop and mobile: open Settings, then Connectors, then Add custom connector, enter the server address and select Connect. ChatGPT connects the same way in developer mode. Smithery, Glama and other directories connect through the same sign-in.
2. Or use a token. In Settings, create a token and copy the secret (it is shown exactly once):
Cursor, VS Code and any other MCP client take the generic remote config:
A client that only speaks stdio can bridge with npx mcp-remote https://mcp.snipzr.com/mcp --header "Authorization: Bearer YOUR_TOKEN".
3. Confirm it worked. Ask your assistant to list the Snipzr tools. Thirteen names starting with snipzr_ means you are connected.
| Tool | What it does | Scope |
|---|---|---|
snipzr_create_link | Create one short link: destination, optional domain, back-half, title | links:write |
snipzr_create_links_batch | Create up to 100 short links in one call | links:write |
snipzr_list_links | List links with text, domain and campaign filters | links:read |
snipzr_get_link | Read one link by its back-half | links:read |
snipzr_update_link | Change a destination or title | links:write |
snipzr_delete_link | Delete a link, with a confirmation gate | links:write |
snipzr_get_link_stats | Per-day clicks and scans, bots counted separately | stats:read |
snipzr_get_campaign_report | Clicks rolled up by campaign | stats:read |
snipzr_get_usage | Plan, remaining allowance and reset date | usage:read |
snipzr_list_domains | Your short domains and their status | domains:read |
snipzr_get_qr_code | A scannable QR code (PNG, JPEG or SVG) with colors, a caption and a logo, plus a signed download URL | links:read |
snipzr_build_utm_link | Compose a UTM link that follows your workspace's conventions | links:read |
snipzr_list_workspaces | The workspaces this connection can act in, with the ids to pass as workspaceId | any |
Every tool carries read-only or destructive annotations, so well-behaved clients ask before a delete and never before a read, and every tool publishes an output schema and returns structured content beside the text, so agents read fields instead of parsing prose.
links:read, links:write, stats:read, usage:read, domains:read. A call outside the token's scopes is refused with a message that names the missing scope.workspaceId argument of any tool (snipzr_list_workspaces lists them). Access follows your live membership there, checked on every call: leave a workspace and the token loses it instantly; get re-added and it resumes at your new role.SCOPE_REQUIRED, NOT_GRANTED, GRANT_INACTIVE, GRANT_REDUCED, PLAN_REQUIRED) so an agent can explain and self-correct.snipzr_get_usage reports both before bulk work.Questions and problem reports are welcome in this repository's issues, or at support@snipzr.com. Please never post a token secret; if one leaks, revoke or roll it in Settings first.