Spec Driven Development
MCP Server that facilitates spec-driven development workflows, not just Vibe Coding.
Description from the maker (original English)
Capture screenshots, detect visual regressions between page versions, and analyze with AI.
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.
SNAPSHOT_SITE_API_KEYValues marked secret are keys to your account. Never share them or paste them into a chat.
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://mcp.snapshot-site.com/mcpOpen Claude Desktop → Settings → Developer → Edit Config, and paste this into claude_desktop_config.json
{
"mcpServers": {
"screenshots": {
"command": "npx",
"args": [
"-y",
"@snapshot-site/mcp"
],
"env": {
"SNAPSHOT_SITE_API_KEY": "your-value"
}
}
}
}Run this in your terminal
claude mcp add --transport http screenshots https://mcp.snapshot-site.com/mcpclaude mcp add screenshots --env SNAPSHOT_SITE_API_KEY=your-value -- npx -y @snapshot-site/mcpFill in the empty values before running it.
Paste into .cursor/mcp.json in your project, or ~/.cursor/mcp.json
{
"mcpServers": {
"screenshots": {
"url": "https://mcp.snapshot-site.com/mcp"
}
}
}Paste into .cursor/mcp.json in your project, or ~/.cursor/mcp.json
{
"mcpServers": {
"screenshots": {
"command": "npx",
"args": [
"-y",
"@snapshot-site/mcp"
],
"env": {
"SNAPSHOT_SITE_API_KEY": "your-value"
}
}
}
}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": {
"screenshots": {
"type": "http",
"url": "https://mcp.snapshot-site.com/mcp"
}
}
}Paste into .vscode/mcp.json in your project
{
"servers": {
"screenshots": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"@snapshot-site/mcp"
],
"env": {
"SNAPSHOT_SITE_API_KEY": "your-value"
}
}
}
}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://mcp.snapshot-site.com/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 8, 2026
The registry name was verified against the domain snapshot-site.com (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.
You need to fill in: SNAPSHOT_SITE_API_KEY
Either way works: install it on your computer or connect to the cloud version.
The GitHub repo last changed –, and the latest version (0.1.5) reached the registry 3 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, Tool Guardian
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.
Too many MCP servers eating your context? A router that loads each tool's schema only when needed.
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.
Official MCP server for the Snapshot Site API
screenshotanalyzecompareThese tools are annotated for MCP clients as:
They also include richer titles, category metadata, and example intents to improve tool selection in Claude Desktop and Cursor.
Hosted, with OAuth — point your client at https://mcp.snapshot-site.com/mcp and sign in. There is no local process to run and no API key in your client config; the server resolves your account from the OAuth session.
Local, with an API key — run the package yourself over stdio and provide SNAPSHOT_SITE_API_KEY. The server calls the Snapshot Site API directly.
Either way, calls count against the same plan quota as direct API calls. Nothing is metered differently because it came through MCP.
The hosted server needs no credential — you sign in through OAuth and the server resolves your account from the session. Skip this section unless you run the server yourself over stdio.
For local stdio mode you need a Snapshot Site API key.
SNAPSHOT_SITE_API_KEY environment variable.SNAPSHOT_SITE_BASE_URL at its default (https://api.prod.ss.snapshot-site.com) unless you're pointed at a self-hosted or staging instance.Requires Node.js 20.9 or later. Built against @modelcontextprotocol/sdk v1.28 and tested with Claude Desktop and Cursor over stdio, and with any client that speaks the streamable HTTP transport against the hosted endpoint.
Not sure which options to use? Preview a capture and its parameters in the Screenshot console playground before wiring them into a tool call.
What a client does when connecting to the hosted server:
1. Discovery
Claude -> GET https://mcp.snapshot-site.com/.well-known/oauth-protected-resource
MCP -> responds: authorization_servers = https://mcp.snapshot-site.com
2. Authorization
Claude -> must know client_id -> opens: https://mcp.snapshot-site.com/oauth/v2/authorize ?client_id=... &redirect_uri=https://claude.ai/api/mcp/auth_callback &response_type=code &code_challenge=...
3. Token
Claude -> receives an access token -> calls the MCP server: POST https://mcp.snapshot-site.com/ Authorization: Bearer
The server then validates the token against the issuer and resolves the account it belongs to before running the tool.
curl -s https://mcp.snapshot-site.com/.well-known/oauth-protected-resource | jq
curl -s https://mcp.snapshot-site.com/.well-known/openid-configuration | jq curl -i https://mcp.snapshot-site.com/ curl -i -X POST https://mcp.snapshot-site.com/mcp -H 'content-type: application/json' --data '{}'
client_id is only needed for the authorization step.client_id beforehand, enter it manually in the connector UI.preferred_client_id in /.well-known/oauth-protected-resource.client_id.export SNAPSHOT_SITE_API_KEY=ss_live_xxx
export SNAPSHOT_SITE_BASE_URL=https://api.prod.ss.snapshot-site.com
pnpm install
pnpm run build
export SNAPSHOT_SITE_API_KEY=ss_live_xxx
snapshot-site-mcp
{
"mcpServers": { "snapshot-site": { "command": "node", "args": ["/absolute/path/to/snapshot-site-mcp/build/server.js"], "env": { "SNAPSHOT_SITE_API_KEY": "ss_live_xxx", "SNAPSHOT_SITE_BASE_URL": "https://api.prod.ss.snapshot-site.com" } } } }
{
"mcpServers": { "snapshot-site": { "command": "node", "args": ["/absolute/path/to/snapshot-site-mcp/build/server.js"], "env": { "SNAPSHOT_SITE_API_KEY": "ss_live_xxx" } } } }
This package also supports a hosted MCP endpoint for clients using mcp-remote.
Start the HTTP server:
pnpm start:http
or:
npx snapshot-site-mcp-http
Environment variables:
export PORT=3000
export HOST=0.0.0.0 export MCP_PATH=/mcp export HEALTH_PATH=/healthz export MCP_ALLOWED_HOSTS=mcp.snapshot-site.com export SNAPSHOT_SITE_BASE_URL=https://api.prod.ss.snapshot-site.com
Remote client configuration with direct API key header:
{
"mcpServers": { "Snapshot Site MCP": { "command": "npx", "args": [ "mcp-remote", "https://mcp.snapshot-site.com/mcp", "--header", "x-snapshotsiteapi-key: ss_live_xxx" ] } } }
The hosted server is stateless. Each request authenticates with x-snapshotsiteapi-key, which makes the service safe to run on multiple replicas without session affinity.
The remote HTTP server also supports OAuth bearer tokens validated against an OIDC issuer such as Zitadel.
Environment variables:
export OIDC_ISSUER_URL=https://auth.snapshot-site.com
export OIDC_AUDIENCE=snapshot-site-mcp export OIDC_REQUIRED_SCOPE=claudeai export OIDC_DISCOVERY_CLIENT_ID=366546620977775166 export RESOURCE_SERVER_URL=https://mcp.snapshot-site.com export ALLOW_API_KEY_AUTH=false export SNAPSHOT_SITE_API_KEY=ss_server_side_xxx
In bearer-token mode, the MCP server validates the incoming access token against the issuer JWKS and then uses the server-side Snapshot Site API key to call the backend API.
It also exposes and proxies:
GET /.well-known/oauth-protected-resource
GET /.well-known/openid-configuration GET/POST /oauth/v2/* GET /ui/* GET/POST /oauth/register
so MCP clients can discover the authorization server metadata automatically.
When OIDC_DISCOVERY_CLIENT_ID is set, the protected resource metadata also includes:
{
"resource_name": "Snapshot Site MCP", "preferred_client_id": "366546620977775166", "oauth_client_metadata": { "client_id": "366546620977775166", "token_endpoint_auth_method": "none" } }
This is an experimental compatibility hint for clients that can infer the OAuth public client automatically. Manual client_id entry remains the reliable fallback.
The MCP server also exposes a lightweight registration_endpoint compatibility shim at:
POST https://mcp.snapshot-site.com/oauth/register
This shim currently returns the preconfigured public PKCE client instead of provisioning a brand-new Zitadel client per installation. It validates and reflects the redirect_uris requested by the client, as long as they are valid HTTPS URLs. It is meant to improve compatibility with clients that expect DCR-style discovery, while keeping the existing manual flow as fallback.