pluggedin_ask_knowledge_base
Ask questions and get AI-generated answers from your knowledge base. Returns structured JSON with answer, document sources, and metadata.
LABSCO SUMMARY
A proxy that also remembers — which is what makes the aggregation worth the extra hop.
Plenty of proxies put many servers behind one connection; what is unusual here is the state that survives the session: a document library you can ask questions of, memory sessions that compress into a summary, and a clipboard entries can be pushed to and popped from as a pipeline between tools. The trade is a dependency: everything routes through the plugged.in app, so its account holds your server configuration and its API key gates the whole connection.
What it is
A proxy that aggregates the MCP servers configured in the plugged.in app behind a single connection, routing calls to the right one. Alongside the proxying it provides its own capabilities: a document library with RAG, session memory, notifications and a clipboard.
What you get
- pluggedin_discover_tools refreshes the tool, resource and template catalogue for the configured servers, for one server or all of them, downstream servers reachable over STDIO, SSE or Streamable HTTP
- pluggedin_ask_knowledge_base answers a question from your document library and returns the answer with its document sources
- Documents created, listed, searched, updated and read: search returns metadata and a snippet, and pluggedin_get_document returns the full content and optionally its version history
- A clipboard shared between tools: entries set by name or pushed to an auto-incrementing index, popped LIFO, listed, fetched and deleted, with a content type, an encoding, a TTL and a visibility per entry, up to 2MB each
- Memory sessions: start a session, record observations with a type and an outcome as work proceeds, and end it to trigger a compressed session summary
- pluggedin_memory_search returns lightweight semantic matches and pluggedin_memory_details fetches the full content of the ones you pick — search first, expand second
- Notifications sent with a severity and optional email delivery, listed with an unread filter, marked done and deleted
- pluggedin_setup explains configuration and needs no API key, which is what makes the first connection debuggable
Requirements
An account on the plugged.in app and an API key from it, passed as PLUGGEDIN_API_KEY or --pluggedin-api-key; PLUGGEDIN_API_BASE_URL points at a self-hosted instance instead of https://plugged.in. It runs as npx -y @pluggedin/pluggedin-mcp-proxy@latest, or from Docker. Default transport is stdio; --transport streamable-http serves over HTTP with --port, --require-api-auth and --stateless as the options around it. Each downstream server still needs its own credentials, configured in the plugged.in app rather than here.
Setup effort
One command plus a key — npx -y @pluggedin/pluggedin-mcp-proxy@latest --pluggedin-api-key <YOUR_PLUGGEDIN_API_KEY>, then supply credentials
USE CASES
01
Running several MCP servers behind one client connection
Configure them in the app, refresh the catalogue, and let the proxy route each call to the server that owns it.
02
Grounding answers in your own documents
Ask the knowledge base, then open the specific document behind the answer in full when it matters.
03
Keeping what an assistant produced
Save the output as a document with tags and category, and append to it later rather than starting a new one each time.
04
Carrying context across sessions
Start a memory session, record observations as the work goes, end it to get a compressed summary, and search that memory next time.
05
Passing a value from one tool to the next
Push results onto the indexed clipboard and pop them in order, or set a named entry with a TTL for something that should expire.
06
Being told when a long job finishes
Send a notification with a severity, optionally by email, and clear it once it has been dealt with.
23 tools
Counted once with only startup setup. Other versions, modes or optional credentials may expose a different list.
pluggedin_clipboard_delete
Delete clipboard entries by name, index, or clear all entries.
pluggedin_clipboard_get
Get clipboard entries. Specify name or idx for a single entry, or omit both to list all entries with pagination.
pluggedin_clipboard_list
List all clipboard entries with metadata. Image values are truncated to first 1000 chars for preview.
pluggedin_clipboard_pop
Pop the highest-indexed entry from the clipboard (LIFO behavior). Returns the entry value and removes it.
pluggedin_clipboard_push
Push a value to the indexed clipboard with auto-incrementing index. Useful for building ordered pipelines or stack-like operations.
pluggedin_clipboard_set
Set a clipboard entry by name (semantic key) or index (array-like). Named entries are upserted; indexed entries fail if index exists. Max 2MB per entry.
pluggedin_create_document
Create and save AI-generated documents to the user's library in Plugged.in (requires API key)
pluggedin_delete_notification
Delete a notification from the Plugged.in system
pluggedin_discover_tools
Triggers discovery of tools (and resources/templates) for configured MCP servers in the Pluggedin App.
pluggedin_get_document
Retrieve a specific document's full content by ID. Use this after pluggedin_search_documents to get the complete content of documents you found. Set includeContent=true to get the full text.
pluggedin_list_documents
List documents with filtering options from the user's library (requires API key)
pluggedin_list_notifications
List notifications from the Plugged.in system with optional filters for unread only and result limit
pluggedin_mark_notification_done
Mark a notification as done in the Plugged.in system
pluggedin_memory_details
Get full details for selected memories (progressive disclosure Layer 3). Use after pluggedin_memory_search to retrieve complete content for specific memories.
pluggedin_memory_observe
Record an observation during a memory session. Observations are classified by the Analytics Agent into memory ring types (procedures, practice, longterm, shocks).
pluggedin_memory_search
Search memories using semantic similarity. Returns lightweight results (50-150 tokens each) for token-efficient progressive disclosure. Use pluggedin_memory_details for full content.
pluggedin_memory_session_end
End a memory session and trigger Z-report generation (AI-compressed session summary).
pluggedin_memory_session_start
Start a new memory session to begin capturing observations. Returns a session UUID and memory_session_id for subsequent operations.
pluggedin_search_documents
Search for specific documents in your library. Returns document metadata (ID, title, snippet). To retrieve full content, use pluggedin_get_document with the returned document ID.
pluggedin_send_notification
Send custom notifications through the Plugged.in system with optional email delivery. You can provide a custom title or let the system use a localized default.
pluggedin_setup
Get started with Plugged.in MCP - shows setup instructions and API key configuration (no API key required)
pluggedin_update_document
Update or append to an existing AI-generated document (requires API key)
BEFORE YOU INSTALL
What to know first
Tools you get
23 tools when we connected. This server is a proxy: tools from the MCP servers connected in your plugged.in workspace are added on top of these 23 built-in tools.
Requires
- npx on PATH
- Node.js 18+
- PLUGGEDIN_API_KEY in your environment
Where it runs
Both documented clients run the server on your machine.
CONTEXT COST
Moderate Standing cost
Rated by the size of the 23 tool definitions that ride along on every turn, not by how many there are
INSTALLATION
Connect your client
Server setup
TERMINAL
npx -y @pluggedin/pluggedin-mcp-proxy@latest --pluggedin-api-key <YOUR_PLUGGEDIN_API_KEY>MCP Clients
Claude Desktop, Cline
JSON
{
"mcpServers": {
"pluggedin": {
"command": "npx",
"args": [
"-y",
"@pluggedin/pluggedin-mcp-proxy@latest"
],
"env": {
"PLUGGEDIN_API_KEY": "<YOUR_PLUGGEDIN_API_KEY>"
}
}
}
}
