How do I use AnythingLLM MCP Server?

1. Click on "Deploy Server". 2. Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state. 3. In the chat, type @ followed by the MCP server name and your instructions, e.g., "@AnythingLLM MCP Server search for documents about machine learning" That's it! The server will respond to your query, and you can continue using it as needed. Here is a step-by-step guide with screenshots.

AnythingLLM MCP Server

by andreperez

Overview Schema Source Related Servers Discussions Admin

Python

Hybrid

AnythingLLM MCP Server

A Model Context Protocol (MCP) server that lets MCP-compatible clients (VS Code, Claude Desktop, Cursor, and others) interact with AnythingLLM. Available tools:

Workspace management
Chat and thread operations
Document operations and embeddings
Vector search
System and model inspection

Requirements

Python 3.10+ — verify: python --version or python3 --version. Download if not installed.
uv (Python package manager) — verify: uv --version. Install instructions.
AnythingLLM running and reachable — open http://localhost:3001 in a browser to confirm. If it runs in Docker, make sure the port is published (e.g., -p 3001:3001).
AnythingLLM API key — create one in the AnythingLLM web interface (look for API Keys or Developer API in Settings). Copy the key and store it securely; it is shown only once.

Related MCP server: AnythingLLM Control Plane MCP Server

Installation

git clone https://github.com/andreperez/anythingllm-mcp.git
cd anythingllm-mcp
uv sync

uv sync reads pyproject.toml and installs all required dependencies into a local virtual environment. You do not need to create or activate the environment manually — uv run handles that automatically.

Configuration

The server reads configuration from these environment variables at startup:

ANYTHINGLLM_BASE_URL: Base URL of your AnythingLLM instance. Defaults to http://localhost:3001 if not set.
ANYTHINGLLM_API_KEY: API key for authenticating with AnythingLLM. This value is required.

The server does not load .env files by itself. You must either:

Set the variables in the shell before starting the server.
Use a tool that loads .env files for you.
Pass the variables directly in your MCP client configuration.

Option 1: Set variables in your current shell

PowerShell:

$env:ANYTHINGLLM_BASE_URL = "http://localhost:3001"
$env:ANYTHINGLLM_API_KEY = "your_api_key_here"

Bash:

export ANYTHINGLLM_BASE_URL="http://localhost:3001"
export ANYTHINGLLM_API_KEY="your_api_key_here"

Command Prompt (cmd.exe):

set ANYTHINGLLM_BASE_URL=http://localhost:3001
set ANYTHINGLLM_API_KEY=your_api_key_here

Use the host-exposed URL for your deployment. For Docker port mappings like 3002:3001, set ANYTHINGLLM_BASE_URL to http://localhost:3002 because the MCP server connects from the host, not from inside the container.

Option 2: Use a `.env` file with `uv`

Copy the example file:

PowerShell:

Copy-Item .env.example .env

Bash:

cp .env.example .env

Command Prompt (cmd.exe):

copy .env.example .env

Edit .env so it contains your real values:

ANYTHINGLLM_BASE_URL=http://localhost:3001
ANYTHINGLLM_API_KEY=replace_with_your_real_api_key

Then start the server with:

uv run --env-file .env anythingllm-mcp

This tells uv to load variables from .env before starting the server.

Option 3: Pass variables in your MCP client configuration

Most MCP clients accept an env block in their configuration file (see the Client setup examples below). When the env block contains ANYTHINGLLM_API_KEY and ANYTHINGLLM_BASE_URL, you do not need to export variables in the shell or use a .env file.

Client setup

Replace /path/to/anythingllm-mcp in the examples below with the absolute path where you cloned the repository. Examples:

Linux / macOS: /home/youruser/anythingllm-mcp
Windows: C:\\Users\\youruser\\anythingllm-mcp (use double backslashes \\ inside JSON strings)

VS Code

Add to your user or workspace mcp.json. Open the Command Palette (Ctrl+Shift+P on Windows/Linux, Cmd+Shift+P on macOS) and run MCP: Open User Configuration:

{
  "servers": {
    "anythingllm": {
      "type": "stdio",
      "command": "uv",
      "args": [
        "run",
        "--directory",
        "/path/to/anythingllm-mcp",
        "anythingllm-mcp"
      ],
      "env": {
        "ANYTHINGLLM_API_KEY": "${input:anythingllm_api_key}",
        "ANYTHINGLLM_BASE_URL": "http://localhost:3001"
      }
    }
  }
}

VS Code supports ${input:name} to prompt for the API key on each session.
In a multi-root workspace, prefer an absolute path for --directory if the MCP server lives outside the current folder root.

If VS Code shows only a partial tool list after you update the server or switch from another AnythingLLM integration to this local one:

Run MCP: List Servers and confirm anythingllm or your chosen server name starts successfully.
Run MCP: Reset Cached Tools to clear stale tool metadata.
Restart the MCP server from MCP: List Servers.

Claude Desktop

Edit claude_desktop_config.json (Settings → Developer → Edit Config):

{
  "mcpServers": {
    "anythingllm": {
      "command": "uv",
      "args": [
        "run",
        "--directory",
        "/path/to/anythingllm-mcp",
        "anythingllm-mcp"
      ],
      "env": {
        "ANYTHINGLLM_API_KEY": "YOUR_API_KEY",
        "ANYTHINGLLM_BASE_URL": "http://localhost:3001"
      }
    }
  }
}

Cursor

Add to .cursor/mcp.json in your project root, or to the global config at ~/.cursor/mcp.json (Linux/macOS) / %USERPROFILE%\.cursor\mcp.json (Windows):

{
  "mcpServers": {
    "anythingllm": {
      "command": "uv",
      "args": [
        "run",
        "--directory",
        "/path/to/anythingllm-mcp",
        "anythingllm-mcp"
      ],
      "env": {
        "ANYTHINGLLM_API_KEY": "YOUR_API_KEY",
        "ANYTHINGLLM_BASE_URL": "http://localhost:3001"
      }
    }
  }
}

Note: ANYTHINGLLM_BASE_URL defaults to http://localhost:3001, which is the application port inside the container. If you publish Docker as 3002:3001, use http://localhost:3002 in client configuration.

Run (standalone)

If you already exported the variables in your shell:

uv run anythingllm-mcp

If you want to load them from .env at launch time:

uv run --env-file .env anythingllm-mcp

Development

Install test dependencies and run the test suite:

uv sync --extra dev
uv run pytest

Troubleshooting

"API key is missing" or 401 Unauthorized

The server could not authenticate with AnythingLLM. Check that:

ANYTHINGLLM_API_KEY is set in your environment or in the env block of your MCP client configuration.
The key matches an active key in AnythingLLM (Settings → API Keys).
The key has no leading or trailing whitespace.

"Connection refused" or timeout

The server could not reach AnythingLLM. Check that:

AnythingLLM is running (open http://localhost:3001 in a browser).
ANYTHINGLLM_BASE_URL points to the correct host and port.
No firewall or VPN is blocking the connection.

Docker port mismatch

If AnythingLLM runs in Docker with a port mapping like 3002:3001:

Inside the container, the application listens on port 3001.
On the host, you access it on port 3002.

Because the MCP server runs on the host (not inside the container), set:

ANYTHINGLLM_BASE_URL=http://localhost:3002

Always use the host port (the left side of the -p or ports: mapping).

MCP server running inside a container (Dev Containers, Docker Compose)

If the MCP server itself runs inside a container (for example, a VS Code Dev Container), localhost refers to that container's own network — not to the host machine. To reach AnythingLLM running on the host or in another container with a published port, use host.docker.internal:

ANYTHINGLLM_BASE_URL=http://host.docker.internal:3001

If both the MCP server and AnythingLLM are containers on the same Docker network (for example, in the same docker-compose.yml), use the service name as hostname instead:

# "anythingllm" is the service name defined in docker-compose.yml
ANYTHINGLLM_BASE_URL=http://anythingllm:3001

In this case, use the container port (3001), not the host-published port.

host.docker.internal is supported on Docker Desktop (Windows and macOS) and on Docker Engine 20.10+ for Linux (requires --add-host=host.docker.internal:host-gateway or the equivalent extra_hosts in Compose).

Security notes

Never commit real API keys.
Use environment variables for secrets.
Use least-privileged AnythingLLM API tokens when possible.

License

MIT

Available Tools

34 tools

anythingllm_chatA

Send a message to a workspace and get a response.

Mode 'chat' uses document context + conversation history. Mode 'query' uses only document context (no history).

Args: slug: Workspace slug message: Message to send mode: 'chat' (context + history) or 'query' (documents only)

Inputs: slug (required), message (required), mode. Returns: result

anythingllm_chat_in_threadB

Send a message within a specific workspace thread.

Args: slug: Workspace slug thread_slug: Thread slug message: Message to send mode: 'chat' or 'query'

Inputs: slug (required), message (required), thread_slug (required), mode. Returns: result

anythingllm_check_authA

Read-onlyIdempotent

Verify that the API key is valid and AnythingLLM is reachable.

Inputs: no named parameters. Returns: result

anythingllm_create_folderC

Create a new document folder in AnythingLLM.

Args: name: Name for the new folder.

Inputs: name (required). Returns: result

anythingllm_create_threadA

Create a new chat thread within a workspace.

Args: slug: Workspace slug name: Optional thread name

Inputs: slug (required), name. Returns: result

anythingllm_create_workspaceB

Create a new workspace.

Args: name: Name for the new workspace

Inputs: name (required). Returns: result

anythingllm_delete_threadA

Destructive

Delete a chat thread from a workspace.

Args: slug: Workspace slug thread_slug: Thread slug to delete

Inputs: slug (required), thread_slug (required). Returns: result

anythingllm_delete_workspaceA

Destructive

Permanently delete a workspace. This action cannot be undone.

Args: slug: Workspace slug to delete

Inputs: slug (required). Returns: result

anythingllm_export_chatsA

Read-onlyIdempotent

Export all chat logs from all workspaces.

Inputs: no named parameters. Returns: result

anythingllm_get_accepted_file_typesA

Read-onlyIdempotent

Get the list of file types that AnythingLLM accepts for upload.

Inputs: no named parameters. Returns: result

anythingllm_get_chat_historyA

Read-onlyIdempotent

Get the chat history for a workspace.

Args: slug: Workspace slug

Inputs: slug (required). Returns: result

anythingllm_get_documentA

Read-onlyIdempotent

Get metadata and details for a specific document by its stored name.

The doc_name is the internal document identifier returned by list_documents (e.g. 'custom-documents/myfile.pdf-abc123.json').

Args: doc_name: Document name/path as returned by the documents API.

Inputs: doc_name (required). Returns: result

anythingllm_get_document_metadata_schemaA

Read-onlyIdempotent

Get the metadata schema that AnythingLLM uses for documents.

Inputs: no named parameters. Returns: result

anythingllm_get_system_settingsA

Read-onlyIdempotent

Get AnythingLLM system settings (LLM provider, vector DB, embeddings, etc.).

Inputs: no named parameters. Returns: result

anythingllm_get_thread_chatsA

Read-onlyIdempotent

Get chat history for a specific thread in a workspace.

Args: slug: Workspace slug thread_slug: Thread slug

Inputs: slug (required), thread_slug (required). Returns: result

anythingllm_get_vector_countA

Read-onlyIdempotent

Get the total number of vectors stored in the system.

Inputs: no named parameters. Returns: result

anythingllm_get_workspaceA

Read-onlyIdempotent

Get detailed information about a specific workspace.

Args: slug: Workspace slug (e.g. 'papers', 'lands')

Inputs: slug (required). Returns: result

anythingllm_list_documentsA

Read-onlyIdempotent

List all uploaded documents across all folders.

Inputs: no named parameters. Returns: result

anythingllm_list_documents_in_folderA

Read-onlyIdempotent

List all documents inside a specific folder.

Args: folder_name: Folder name to list documents from.

Inputs: folder_name (required). Returns: result

anythingllm_list_embedsA

Read-onlyIdempotent

List all embed configurations (public chat widgets).

Inputs: no named parameters. Returns: result

anythingllm_list_modelsA

Read-onlyIdempotent

List available models via the OpenAI-compatible endpoint.

Inputs: no named parameters. Returns: result

anythingllm_list_workspacesA

Read-onlyIdempotent

List all workspaces in the AnythingLLM instance with their slugs, settings, and thread info.

Inputs: no named parameters. Returns: result

anythingllm_move_filesA

Move documents between folders.

Each entry in 'files' must have 'from' and 'to' keys with document paths as returned by list_documents (e.g. 'custom-documents/file.pdf-uuid.json').

Args: files: List of move operations, each with 'from' and 'to' path strings. Example: [{"from": "custom-documents/a.pdf-uuid.json", "to": "custom-documents/my-folder/a.pdf-uuid.json"}]

Inputs: files (required). Returns: result

anythingllm_remove_documentsA

Destructive

Permanently delete one or more documents from the system.

This removes the documents from AnythingLLM storage entirely. They will also be removed from any workspaces that had them embedded.

Args: names: List of document names/paths as returned by list_documents (e.g. ['custom-documents/myfile.pdf-abc123.json']).

Inputs: names (required). Returns: result

anythingllm_remove_folderA

Destructive

Permanently delete a document folder and all its contents.

Args: name: Folder name to delete.

Inputs: name (required). Returns: result

anythingllm_searchA

Read-onlyIdempotent

Search for relevant document chunks within a workspace using vector similarity.

Args: slug: Workspace slug to search in query: Search query top_n: Number of results (1-20, default 4) score_threshold: Similarity score threshold (0-1). Lower = more results.

Inputs: slug (required), query (required), top_n, score_threshold. Returns: result

anythingllm_update_embeddingsA

Idempotent

Embed or un-embed documents in a workspace (equivalent to "Save and Embed" in the UI).

This is the final step after uploading documents to AnythingLLM storage. Typical workflow:

Upload a file -> upload_file / upload_file_to_folder / upload_link / upload_raw_text
List documents -> list_documents / list_documents_in_folder (get the doc paths)
Embed into workspace -> THIS TOOL with 'adds' containing the doc paths from step 2

Once embedded, the document's content is vectorized and available for RAG queries in the workspace. Removing (via 'deletes') un-embeds the document from the workspace but does NOT delete it from storage.

Args: slug: Workspace slug where documents will be embedded. adds: Document paths to embed, as returned by list_documents (e.g. ['custom-documents/Pine Script/file.pine-abc123.json']). deletes: Document paths to un-embed (remove from workspace only, the file remains in AnythingLLM storage).

Inputs: slug (required), adds, deletes. Returns: result

anythingllm_update_pinA

Idempotent

Pin or unpin a document in a workspace.

Pinned documents are always included in the LLM context for every query, regardless of vector similarity results.

Args: slug: Workspace slug. doc_path: Document path as returned by list_documents (e.g. 'custom-documents/myfile.pdf-abc123.json'). pinned: True to pin, False to unpin.

Inputs: slug (required), pinned (required), doc_path (required). Returns: result

anythingllm_update_threadA

Idempotent

Update the name of an existing chat thread.

Args: slug: Workspace slug. thread_slug: Thread slug to update. name: New name for the thread.

Inputs: name (required), slug (required), thread_slug (required). Returns: result

anythingllm_update_workspaceA

Idempotent

Update workspace settings (name, temperature, prompt, similarity threshold, etc.).

Args: slug: Workspace slug to update name: New workspace name openAiTemp: LLM temperature (0.0-1.0) openAiHistory: Chat history length (0-100) openAiPrompt: System prompt override similarityThreshold: Similarity threshold (0.0-1.0) topN: Top N results for context (1-20) chatMode: Chat mode: 'chat' or 'query'

Inputs: slug (required), name, topN, chatMode, +4 more. Returns: result

anythingllm_upload_fileC

Upload a file from the local filesystem to AnythingLLM.

Args: file_path: Absolute path to the file on the local filesystem.

Inputs: file_path (required). Returns: result

anythingllm_upload_file_to_folderA

Upload a file from the local filesystem into a specific document folder in AnythingLLM.

Args: file_path: Absolute path to the file on the local filesystem. folder_name: Target folder name in AnythingLLM (must already exist).

Inputs: file_path (required), folder_name (required). Returns: result

anythingllm_upload_linkA

Scrape a web page and upload it as a document to AnythingLLM.

Args: link: URL to scrape (must start with http:// or https://)

Inputs: link (required). Returns: result

anythingllm_upload_raw_textA

Upload raw text content as a document to AnythingLLM.

Args: text_content: Raw text content to upload title: Document title

Inputs: title (required), text_content (required). Returns: result

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

Deploy Server

One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

TDQS

A3.8/5.0

Scored 2026-08-07 01:45 across 34 tools

Disambiguation5/5

Every tool targets a distinct resource and action. Upload variants differ by input source (file, link, raw text, folder), chat tools differ by thread versus workspace, and list_documents vs list_documents_in_folder are clearly scoped. No two tools appear to do the same thing.

Naming Consistency5/5

All tool names follow a consistent 'anythingllm_verb_noun' pattern using lower_snake_case. Examples include list_workspaces, create_thread, upload_file, get_vector_count, and remove_documents. The prefix and verb ordering are uniform, making the toolset predictable.

Tool Count2/5

34 tools is well above the rubric's 'too many' threshold of 25. While the server covers a wide range of application features, the sheer volume could overwhelm agents and suggests the surface is not tightly scoped. Many system/settings/export tools could potentially be consolidated.

Completeness5/5

The toolset provides comprehensive lifecycle coverage for workspaces, threads, documents, folders, embeddings, and chat. It includes auth validation, upload variants, pinning, search, vector counts, export, and embed listing—leaving no obvious gaps in the core AnythingLLM workflow.

Maintenance

ActivitySlowing

ResponsivenessNo recent community issues

Resources

Related MCP Connectors

mcp-server
com.niftypm
Nifty's MCP server — exposes tasks, projects, messages, and files as tools for AI agents.
Project Management Communication File Systems
2
https://mcp.niftypm.com/mcp
Flowra
io.github.flowradev
MCP server for your apps' tools and custom tools, plus hosted AI agents and approval-gated workflows
Agent Orchestration Autonomous Agents
11
https://mcp.flowra.dev/mcp
mcp
com.skopik
Your org's AI agents, tasks, runs, search, and brain files as MCP tools and resources.
Agent Orchestration Knowledge & Memory Search
2
https://api.skopik.com/mcp
Backtest360A
com.backtest360
MCP server exposing the Backtest360 engine API as tools for AI agents.
Finance Research & Data
202
https://mcp.backtest360.com/mcp

Related MCP Servers

anythingllm-mcp
RAG Systems Search
moliver28
A
license
C
quality
D
maintenance
MCP server for AnythingLLM that enables AI document chat platform interaction with tools for workspaces, chat, documents, threads, and system operations.
Updated 6 months ago (2026-05-02 05:15 UTC)
17
MIT
AnythingLLM Control Plane MCP
RAG Systems AI & Machine Learning
TheNovaNodes
A
license
A
quality
B
maintenance
Enables managing and controlling AnythingLLM instances through MCP, providing tools to list, create, and delete workspaces, retrieve system environment settings, and get vector counts.
Updated 2 months ago (2026-09-05 21:32 UTC)
5
1
MIT
MSIM
AI & Machine Learning Agent Orchestration
giancbaring
A
license
Not graded
quality
C
maintenance
Enables AnythingLLM to be controlled by any MCP-compatible client via HTTP, SSE, or WebSocket, with automatic workspace selection and background operation.
Updated 2 months ago (2026-08-27 09:43 UTC)
MIT
anythingllm-mcp
Search RAG Systems
DezGDNY
A
license
A
quality
C
maintenance
Enables AI assistants to search and ask questions about documents stored in AnythingLLM workspaces through MCP tools.
Updated 2 months ago (2026-09-09 01:10 UTC)
4
MIT

AnythingLLM MCP Server

Requirements

Installation

Configuration

Option 1: Set variables in your current shell

Option 2: Use a .env file with uv

Option 3: Pass variables in your MCP client configuration

Client setup

VS Code

Claude Desktop

Cursor

Run (standalone)

Development

Troubleshooting

"API key is missing" or 401 Unauthorized

"Connection refused" or timeout

Docker port mismatch

MCP server running inside a container (Dev Containers, Docker Compose)

Security notes

License

Available Tools

Tool Schema Changelog

TDQS

Maintenance

Resources

Related MCP Connectors

Related MCP Servers

anythingllm-mcp

AnythingLLM Control Plane MCP

MSIM

anythingllm-mcp

Option 2: Use a `.env` file with `uv`