Spec Driven Development
MCP Server that facilitates spec-driven development workflows, not just Vibe Coding.
Description from the maker (original English)
Test the idea before you build it. Rapidly works inside your AI agent.
These three are never merged into one score: opinions, measurements and popularity are different things.
We have not written a plain-language summary for this one yet. Below are the facts we read from the registry.
Pick the AI app you use, then copy the settings.
You need
Which app do you use?
Pick your app above. We remember it, so next time the button says “Install in …”.
Works in Claude on the web, desktop and mobile:
https://www.rapidly.co/mcpRun this in your terminal
claude mcp add --transport http rapidly https://www.rapidly.co/mcpPaste into .cursor/mcp.json in your project, or ~/.cursor/mcp.json
{
"mcpServers": {
"rapidly": {
"url": "https://www.rapidly.co/mcp"
}
}
}Your browser asks to open the app, then the app shows the settings and asks you to confirm. Key values are left empty for you to fill in.
Paste into .vscode/mcp.json in your project
{
"servers": {
"rapidly": {
"type": "http",
"url": "https://www.rapidly.co/mcp"
}
}
}Your browser asks to open the app, then the app shows the settings and asks you to confirm. Key values are left empty for you to fill in.
ChatGPT connects only to servers on the internet:
https://www.rapidly.co/mcpMenu names can differ slightly by plan and over time.
Your choice stays on this device in a cookie; no account needed.
These settings are generated from the maker’s registry entry. If the README says otherwise, follow the README.
We have not read this server’s tool list yet; we check endpoints in rotation.
We check information anyone can see; we have not read the code. Checks that do not apply are left out of the count. · Checked Nov 6, 2026
The registry name was verified against the domain rapidly.co (the registry checks domain ownership)
We have not read this repo from GitHub yet
We have not read this repo from GitHub yet
Not checked yet; we check endpoints in rotation
Not checked yet; we check endpoints in rotation
Not checked yet; we read the tool list when we check the endpoint
Compared with every active server in this category in the registry. Stars count only when the publisher owns the repo.
It passed 1 of 1 applicable basic checks. We have not read the code. Before adding a key or token, see what access the server asks for and use a key with only the permissions you need.
The registry does not list a required key, though some servers ask you to sign in on first use.
It runs in the maker’s cloud. Connect over the internet with nothing to install, but what you send passes through the maker’s servers.
The GitHub repo last changed –, and the latest version (2.0.0) reached the registry 2 months ago.
Apps that support MCP, such as Claude Desktop, Claude Code, Cursor and VS Code. It connects by URL, so apps that accept remote MCP servers work too, such as ChatGPT in developer mode.
In the same category: Spec Driven Development, Remote Desktop Commander, topo^alpin
Servers in the same category from other makers, sorted by stars.
Reading a card:★ GitHub starsiThree pillars: reviews · checks · usageiUpdated = latest version published to the registryi
MCP Server that facilitates spec-driven development workflows, not just Vibe Coding.
Remote Desktop Commander: use files, terminal and processes on your own computers from any AI app.
Avalanche bulletins and mountain weather forecasts for any place in the Alps.
Copied from the maker’s repo as written. PopNx MCP has not reviewed or endorsed it. HTML, images and scripts are removed. Links lead off this site.
Rapidly works inside your AI agent. Bring an idea you already have, or ask it to generate one from a company website. In about five minutes, you get a build prompt for the smallest Pretotyping experiment worth running.
Rapidly builds the Lean Canvas, turns the riskiest assumption into a measurable hypothesis, researches a pass mark and designs practical experiments. Your agent builds the test. Real customers give you the evidence.
Rapidly MCP is generally available.
Version 2.0.0 is published in the official MCP Registry.
Works with Claude, Codex, ChatGPT and Hermes. OpenClaw, Grokbot and other agents can also use Rapidly if they support remote MCP.
Connect Rapidly · Let your agent set it up · Manual setup
A suburban independent bookshop wants to grow weekday foot traffic and build a younger, regular customer base.
Rapidly generated three ideas. We chose the Third-Place Pass: a $19 monthly membership with a guaranteed two-hour weekday seat, a rotating perk and a members' shelf.
Rapidly built the Lean Canvas, turned the riskiest demand assumption into a measurable hypothesis and proposed five Pretotyping experiments. The simplest started with staff asking warm, local customers to pay. The next used a counter card and landing page, with a proposed pass mark of 15%.
See the complete Rapidly run · Open the customer-facing landing page
This was a demonstration for a hypothetical bookshop. The pass mark was a proposal to test, not a measured result. No customer results are claimed.
You can. We do. The answer will probably look reasonable.
Rapidly is useful because the method is already there. For one idea, the core run:
Your agent handles that complete run through test_idea. The other experiment options stay saved with the idea, ready to use later. See the seven standard tools and three advanced method tools.
Successful work is also saved with the idea in your Rapidly team instead of disappearing in a chat thread. You can bring the result back later and decide whether to fund it, fix it, stop it or run another test.
The method combines Alberto Savoia's Pretotyping with our own practice through 4,000+ experiments with client teams and 200+ workshops.
The plugin bundles the Rapidly MCP connection with four skills that tell your agent when and how to use it: test an idea, record a test result, choose the next idea, and setup.
In Claude Code:
/plugin marketplace add Exponentially-Platform/Rapidly-MCP
/plugin install rapidly@rapidly
Then run /mcp, choose Rapidly and sign in.
In Codex:
codex plugin marketplace add Exponentially-Platform/Rapidly-MCP
Then open the Plugins Directory in the ChatGPT desktop app or Codex, choose the Rapidly marketplace and install Rapidly. Sign in when Rapidly asks.
In Cursor, the repository carries a Cursor marketplace manifest that points at the same plugin. Install Rapidly from the Cursor marketplace once it is listed, then sign in when Rapidly asks.
The plugin lives in plugins/rapidly. scripts/package-openai-plugin.sh builds the ZIP for OpenAI's plugin directory.
Open rapidly.co/rapidly-mcp and enter your name, email and team name. Rapidly creates the account and team, then shows your personal token once.
The page gives you copy-ready setup for Codex CLI and Claude Code. Follow the steps, start a fresh conversation and you're ready to try an idea.
Connect with OAuth, choose your Rapidly team and approve access. You don't need to create another account or personal token.
The exact steps differ between Codex, Claude and ChatGPT. Use the setup instructions for your agent.
Give your agent this address and ask it to connect Rapidly:
https://www.rapidly.co/mcp
Use OAuth if the agent supports it. Otherwise, follow the agent setup instructions to keep your personal token out of chat and project files.
To generate three ideas and fully test all three in one request:
Ask Rapidly for three ideas for [company], focused on [opportunity], and fully test all three in one go.
Rapidly should save all three ideas, complete each Canvas and hypothesis, save three to five experiment options for each idea, and return one build prompt for each recommended first test.
Use one you already have:
Ask Rapidly to test this idea: [describe the customer, problem and proposed change].
Or ask Rapidly to generate one from a company and its website:
Ask Rapidly to generate [one or more] ideas for [company name] using [website URL]. Focus on [optional customer, problem or opportunity].
Rapidly researches the company, returns grounded ideas with sources and offers to take one through the method. If you're bringing your own idea, add anything you already know, where that information came from and any constraints that matter.
Rapidly works through the method one saved stage at a time. Your agent should continue the same test_idea automatically while Rapidly reports that work is still in progress. Keep the returned project link; if the conversation stops, include that link or project ID when asking Rapidly to continue. You don't need to choose stages, decide whether to save or manage a retry key.
By the end, you should have:
Review the output, ask your agent to build the asset, then put the test in front of real customers.
See more example prompts
We're looking for the points where this becomes confusing or stops being useful.
After your run, please tell us:
Email hello@rapidly.co with the subject Rapidly MCP feedback. Tell us which agent you used and where you stopped. We won't share your idea, feedback or results publicly without asking you first. Please remove tokens and customer data from feedback, logs and screenshots.