Apify MCP Server
MCP setupData and AI
FreeMCP setupData and AI
Give your agent a semantic memory in Qdrant: store notes and code snippets, then find them again by meaning with the official mcp-server-qdrant.
npx workspacetemplates add qdrant-mcp --tool claude-codeWrites 1 file to your project root
{
"mcpServers": {
"qdrant": {
"command": "uvx",
"args": [
"mcp-server-qdrant"
],
"env": {
"QDRANT_URL": "${QDRANT_URL}",
"QDRANT_API_KEY": "${QDRANT_API_KEY}",
"COLLECTION_NAME": "${COLLECTION_NAME}",
"QDRANT_READ_ONLY": "${QDRANT_READ_ONLY}"
}
}
}
}npx workspacetemplates add qdrant-mcp --tool codexWrites 1 file to ~/.codex/
[mcp_servers.qdrant]
command = "uvx"
args = ["mcp-server-qdrant"]
[mcp_servers.qdrant.env]
QDRANT_URL = "${QDRANT_URL}"
QDRANT_API_KEY = "${QDRANT_API_KEY}"
COLLECTION_NAME = "${COLLECTION_NAME}"
QDRANT_READ_ONLY = "${QDRANT_READ_ONLY}"npx workspacetemplates add qdrant-mcp --tool cursorWrites 1 file to .cursor/
{
"mcpServers": {
"qdrant": {
"command": "uvx",
"args": [
"mcp-server-qdrant"
],
"env": {
"QDRANT_URL": "${env:QDRANT_URL}",
"QDRANT_API_KEY": "${env:QDRANT_API_KEY}",
"COLLECTION_NAME": "${env:COLLECTION_NAME}",
"QDRANT_READ_ONLY": "${env:QDRANT_READ_ONLY}"
}
}
}
}npx workspacetemplates add qdrant-mcp --tool gemini-cliWrites 1 file to .gemini/
{
"mcpServers": {
"qdrant": {
"command": "uvx",
"args": [
"mcp-server-qdrant"
],
"env": {
"QDRANT_URL": "$QDRANT_URL",
"QDRANT_API_KEY": "$QDRANT_API_KEY",
"COLLECTION_NAME": "$COLLECTION_NAME",
"QDRANT_READ_ONLY": "$QDRANT_READ_ONLY"
}
}
}
}npx workspacetemplates add qdrant-mcp --tool copilotWrites 1 file to .vscode/
{
"inputs": [
{
"type": "promptString",
"id": "qdrant_url",
"description": "URL of your Qdrant server, for example http://localhost:6333",
"password": true
},
{
"type": "promptString",
"id": "qdrant_api_key",
"description": "API key for Qdrant Cloud; leave unset for a local server",
"password": true
},
{
"type": "promptString",
"id": "collection_name",
"description": "Default collection to store into and search",
"password": true
},
{
"type": "promptString",
"id": "qdrant_read_only",
"description": "Set to true to disable the store tool",
"password": true
}
],
"servers": {
"qdrant": {
"type": "stdio",
"command": "uvx",
"args": [
"mcp-server-qdrant"
],
"env": {
"QDRANT_URL": "${input:qdrant_url}",
"QDRANT_API_KEY": "${input:qdrant_api_key}",
"COLLECTION_NAME": "${input:collection_name}",
"QDRANT_READ_ONLY": "${input:qdrant_read_only}"
}
}
}
}The official Qdrant server exposes two tools. One stores a piece of text with optional metadata in a collection, and the other finds the stored entries closest in meaning to a query. It suits a long-lived memory for an agent, or a searchable index of snippets.
It runs through uvx, so you need uv installed. Point it at a Qdrant instance with a URL and key, or at a local folder for a file-based database. Embeddings are made locally with fastembed by default, so no embedding API key is involved.
Setting QDRANT_READ_ONLY to true removes the store tool and leaves lookup only.
you › Store this: 'Invoices are rounded per line, not per total.' Then find what we decided about rounding.
agent › Stored one entry in the 'decisions' collection. Closest match: 'Invoices are rounded per line, not per total.'MCP setupData and AI
FreeMCP setupData and AI
FreeMCP setupData and AI
FreeMCP setupData and AI
Free