Spec Driven Development
MCP Server that facilitates spec-driven development workflows, not just Vibe Coding.
Description from the maker (original English)
A self-hostable feed of reminders, learnings and work that Claude and ChatGPT can read and write
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://cardlake.app/api/mcpRun this in your terminal
claude mcp add --transport http cardlake https://cardlake.app/api/mcpPaste into .cursor/mcp.json in your project, or ~/.cursor/mcp.json
{
"mcpServers": {
"cardlake": {
"url": "https://cardlake.app/api/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": {
"cardlake": {
"type": "http",
"url": "https://cardlake.app/api/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://cardlake.app/api/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.
The server asks you to sign in before it lists its tools.
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 cardlake.app (the registry checks domain ownership)
We have not read this repo from GitHub yet
We have not read this repo from GitHub yet
Responded in 227 ms (HTTP 401)
Declares OAuth sign-in, so your AI app can take you to log in
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 3 of 3 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 (1.0.0) reached the registry 2 days 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, Reelsaw
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.
Drive a self-hosted Reelsaw: turn long videos into vertical clips, edit, re-render and export them.
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.
A self-hostable feed that Claude and ChatGPT can save to. Reminders, learnings, bookmarks and work follow-ups in one place you own: open source (AGPL), on your own Vercel and Neon accounts.
[Get started](#get-started) · What it does · Looking after your Cardlake · For developers
Don't want to host it yourself? Use Cardlake at cardlake.app.
To run your own: free on Vercel's and Neon's free plans, about 10 minutes, no terminal.
If Vercel shows its dashboard instead, come back and click the button again.
New to Vercel or GitHub? The step-by-step guide has a picture for every step.
The whole deploy, from the button to a first card, in 37 seconds:
https://github.com/user-attachments/assets/8834c580-ecd8-4ebc-847b-f73f6f40099c
Cardlake turns everything you don't want to forget (reminders, articles to read, family updates, bookmarks, notes, and your GitHub issues) into a single feed, instead of a dozen apps you never open.
https://github.com/user-attachments/assets/10d1ce6c-f12a-42bf-aed6-dba16530b934
Notes apps, task apps and read-later apps each hold a piece of what you meant to come back to, and you check none of them. Cardlake is the one place it all lands, and the place your AI writes to when you ask it to remember something. Because your own copy runs on your own Vercel and Neon accounts, what you and your AI save stays with you.
Something you mean to come back to keeps nagging at you until it's written down somewhere you trust. One feed is easier to keep up with than seven apps, and it brings things back to the top before you forget them.
Your Cardlake has one account, yours. There's nobody to administer: no sign-up, no user list, no invites. Sign in with the same email and password on your phone and other computers.
In Vercel, open your project, then Settings and Environment Variables. Add CARDLE_SETUP_OPEN with the value 1, save, and redeploy (Deployments, ⋯, Redeploy). Open /setup on your site and set a new password with the same email: you keep all your cards. Then delete CARDLE_SETUP_OPEN and redeploy again.
Your copy doesn't update by itself. When a new version is out, Cardlake shows Update available at the top; it leads to your Cardlake's /updates page. The first time, click Turn on updates there: GitHub opens with the update file filled in, ready to commit. After that, to get the latest Cardlake, open your copy on GitHub, then Actions, Update Cardlake and Run workflow. Vercel deploys the new version a minute or two later, and you keep all your cards. If you've changed the code yourself, it stops rather than erase your changes. To also hear about new versions by email, open Cardlake on GitHub and choose Watch, Custom, Releases.
Everything else is optional. Cardlake makes its own secret keys and sets up its own database. docs/ENVIRONMENT.md lists the settings you can add later, such as a custom domain, and docs/ADMIN.md covers running it.
git clone https://github.com/cardleforall/cardlakeforall.git
cd cardlakeforall npm ci cp .env.example .env.local # fill in the values -- see docs/ENVIRONMENT.md npm run dev
This starts a local dev server at http://localhost:3000 that serves public/ statically and routes /api/* to the same handler files Vercel uses in production; no separate build step, no vercel dev required. Point DATABASE_URL at a database of your own, never one with real data.
npm ci
npm test
npm run build:vault validates the committed shell and accepts --target for tests. Outside services that feed cards in (currently GitHub Issues) are integrations; docs/INTEGRATIONS.md covers how they work and how to add one. See .env.example and docs/ENVIRONMENT.md for what each environment variable does.
npm test covers the build; the actual test suite (tests/unit/ offline unit tests, tests/sit/ system integration tests against a deployment) is run separately via npm run test:lite / npm run test:full; see tests/README.md for the unit/SIT split and when to run which tier.
public/vault/index.html is a committed, hand-maintained shell (not generated from markdown or any other source), progressively enhanced with plain JS in public/vault/assets/.api/, consolidated per resource area to stay within Vercel's Hobby function-count limit (see .vercelignore and the comments in api/vault.js / api/auth.js)./setup (see docs/ADMIN.md).schema-v2.sql and applies itself on the first request after a deploy.api/mcp.js + api/_lib/mcp-tools.js expose every card type and feed action (snooze, pin, favorite) to any MCP-compatible AI assistant. See docs/MCP_SERVER.md.Deployment is GitHub-to-Vercel: pushes to the connected main branch create production deployments; pull requests create previews. See DEPLOY.md.
Issues are welcome; pull requests aren't accepted. See CONTRIBUTING.md for why, and for setting up and testing your own copy. Please also read the Code of Conduct.
Cardlake is free software under the GNU AGPL-3.0. You can use, change and run it freely. If you run a modified version as a service for other people, you must offer them its source under the same license.
Cardlake™ is a trademark of the project's maintainer. The license covers the code, not the name: run your own copy as you like, but please don't offer a service to other people under the name Cardlake.
Security issues: please report them privately, see SECURITY.md.