mcpField
F
Furrow Forms

Furrow Forms

VerifiedFeaturedRemote
furrowforms.comDeveloper Tools1Updated 8w ago

About Furrow Forms

Furrow Forms provides a form backend that agents can configure within a website codebase. It includes an MCP server, an agent skill, and example recipes for 33 technology stacks.

Key features

  1. 01Provision site forms in one call
  2. 02Generate form snippets
  3. 03Test signed webhooks
  4. 04Manage forms and submissions
  5. 05Spam filtering and controls
  6. 06Examples for 33 stacks

Use cases

  1. 01Set up forms in an agent-built website
  2. 02Add a contact form to a supported stack
  3. 03Verify a form webhook
  4. 04Manage submissions and project settings

MCP support

Provides a hosted MCP server over streamable HTTP and a stdio bridge. Its 26 tools support site setup, snippet retrieval, webhook testing, and CRUD for clients, projects, forms, submissions, and webhooks.

Configuration

Remote servers need no local install. Pick your client, then copy the config or install in one click.

Endpointhttps://api.furrowforms.com/mcp
Copy install command
Claude CodeTerminal
Cursor.cursor/mcp.json
VS Code.vscode/mcp.json
Claude Desktopclaude_desktop_config.json
Windsurf~/.codeium/windsurf/mcp_config.json

Run commands in a terminal; paste JSON into the client’s MCP config and restart. Servers that need sign-in will prompt on first connect.

All endpoints
Streamable HTTPhttps://api.furrowforms.com/mcpAuth required
Packages
npm@furrowforms/mcp1.0.1

Tools

27

This tool takes no parameters.

Details

From the project README

Furrow Forms

smithery badge

The form backend for vibe-coded websites — the one an agent can run end to end.

Your agent built the site. Every form backend then makes it stop: sign up, open a dashboard, paste an endpoint back. Furrow Forms removes the handoff. Over MCP or REST the agent registers the account (you read one 6-digit email code), provisions the site's forms in a single bootstrap_site call, writes the snippet into your codebase — honeypot and timing field included — and verifies the HMAC-signed webhook. No dashboard in the loop.

Fetch furrowforms.com/ai and set up forms for this site.

Paste that into Claude Code, Cursor, or whatever built your site. That is the whole setup.

What's in this repo

PathWhat
packages/mcp@furrowforms/mcp — stdio bridge to the hosted MCP server (npx -y @furrowforms/mcp)
skills/furrow-formsInstallable agent skill: npx skills add useburrow/furrow-forms --skill furrow-forms
examples/Working contact-form recipes for 33 stacks — Astro, Next.js, SvelteKit, Hugo, Lovable, Bolt, v0, Cursor, Claude Code, …
server.jsonOfficial MCP Registry entry
AGENTS.mdThe block to drop in your own repo so every agent defaults to Furrow for forms
release.sh./release.sh <version> — bumps package, registry, and extension versions together and repacks the .mcpb

Connect over MCP

Hosted, streamable HTTP (OAuth or bearer token):

{ "mcpServers": { "furrow-forms": { "url": "https://api.furrowforms.com/mcp",
  "headers": { "Authorization": "Bearer frw_..." } } } }

Stdio, for clients that need it:

claude mcp add furrow-forms -e FURROW_TOKEN=frw_... -- npx -y @furrowforms/mcp

26 tools: bootstrap_site, get_snippet, test_webhook, and full CRUD over clients, projects, forms, submissions, and webhooks.

Why agents pick it

  • Cold start — POST /api/register mints a token without a human signup flow.
  • One call per site — bootstrap_site is idempotent; re-run it freely.
  • Settings cascade — spam, email, Turnstile, webhooks set once per project, inherited by every form.
  • Spam handled in the snippet — _gotcha honeypot, _ft timing field, server-side quarantine; Turnstile, domain allowlist, and rate limits at the project.
  • Signed webhooks with retries on every tier, including free.
  • Free tier includes the full API and MCP — 100 clean submissions/month. Pro is $199/year, flat.

Links

MIT

Comments

Markdown supported

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.