About Flarelink
Flarelink provides an MCP server for AI coding tools building apps with Cloudflare auth, D1, and R2. It supplies stack guidance and code patterns, with optional tools to manage Flarelink projects.
Key features
- 01Cloudflare stack guidance
- 02App scaffolding instructions
- 03Canonical code patterns
- 04SDK and cost references
- 05Optional project management tools
Use cases
- 01Scaffolding a Flarelink application
- 02Applying safe auth, D1, and R2 patterns
- 03Checking Cloudflare cost guidance
- 04Listing projects, databases, or buckets
MCP support
The MCP server exposes stack guides, scaffolding guidance, patterns, SDK and cost references as tools, plus a stack-guide resource. With FLARELINK_API_KEY, it also offers tools to inspect projects, query D1, and list R2 buckets.
Configuration
Local server config is the same for every client. Copy it into your client’s MCP config file.
npx -y @flarelink/mcpFLARELINK_API_KEYSecretOptional. Personal API key from dash.flarelink.dev → API keys. Enables the project-management tools; the knowledge & scaffolding tools work without it.Local servers need a runtime on your machine (e.g. Node.js 18+). Config file locations differ by client; see your client’s docs.
Tools
11Saved tool list · updated 2w ago
This tool takes no parameters.
Details
From the project README@flarelink/mcp
An MCP server that teaches your AI coding tool to build correctly on the Flarelink stack — Cloudflare auth (better-auth + KV sessions), D1, and R2 — from inside Cursor, Claude Code, Windsurf, and anything else that speaks MCP.
It hands the agent the security and cost rules up front (user-scoping, identifier-safe SQL, KV sessions, presigned R2) so the generated code is right the first time, instead of the agent guessing and you debugging.
Install
No install needed — run it with npx. Add it to your tool's MCP config:
Claude Code
claude mcp add flarelink -- npx -y @flarelink/mcp
Cursor / Windsurf / Claude Desktop (mcp.json / config):
{
"mcpServers": {
"flarelink": {
"command": "npx",
"args": ["-y", "@flarelink/mcp"],
"env": { "FLARELINK_API_KEY": "flk_…" }
}
}
}
The FLARELINK_API_KEY is optional. Without it you still get all the
knowledge + scaffolding tools. With it, the management tools (below) can
drive your actual Flarelink projects. Mint a key at
dash.flarelink.dev → API keys.
Tools
| Tool | What it gives the agent |
|---|---|
flarelink_stack_overview | The stack, cardinal rules, and deployment shapes. Read first. |
flarelink_scaffold | How to bootstrap a complete working app (clone / one-click deploy) + file map. |
flarelink_list_patterns | The catalog of canonical code patterns. |
flarelink_get_pattern | Copy-pasteable code for one recipe (auth setup, route guards, identifier-safe D1, R2 upload/presign, wrangler bindings, SDK usage). |
flarelink_sdk_reference | @flarelink/client signatures + return shapes (auth / storage / db). |
flarelink_cost_patterns | Cloudflare cost-optimization guidance, with targeted hints for the feature you describe. |
It also exposes the stack guide as a resource (flarelink://stack-guide) you can attach as context.
Management tools (need FLARELINK_API_KEY)
Drive your actual Flarelink projects from the editor. These route through the Flarelink dashboard API and act exactly as you would in the UI.
| Tool | What it does |
|---|---|
flarelink_whoami | Verify the key; show the user + active connection/project. |
flarelink_list_projects | List projects on the active Cloudflare connection. |
flarelink_list_databases | List the project's D1 databases. |
flarelink_query_database | Run a (parameterized) SQL statement against a project D1. |
flarelink_list_buckets | List the project's R2 buckets. |
Without a key they return a clear "set FLARELINK_API_KEY" message. The key has the same access as your Flarelink login — treat it like a password.
License
MIT. Part of Flarelink — the dashboard for the Cloudflare developer stack.
Comments
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.