Initial commit

.gitignore

@@ -0,0 +1,162 @@
+# Byte-compiled / optimized / DLL files
+__pycache__/
+*.py[cod]
+*$py.class
+
+# C extensions
+*.so
+
+# Distribution / packaging
+.Python
+build/
+develop-eggs/
+dist/
+downloads/
+eggs/
+.eggs/
+lib/
+lib64/
+parts/
+sdist/
+var/
+wheels/
+share/python-wheels/
+*.egg-info/
+.installed.cfg
+*.egg
+MANIFEST
+
+# PyInstaller
+#  Usually these files are written by a python script from a template
+#  before PyInstaller builds the exe, so as to inject date/other infos into it.
+*.manifest
+*.spec
+
+# Installer logs
+pip-log.txt
+pip-delete-this-directory.txt
+
+# Unit test / coverage reports
+htmlcov/
+.tox/
+.nox/
+.coverage
+.coverage.*
+.cache
+nosetests.xml
+coverage.xml
+*.cover
+*.py,cover
+.hypothesis/
+.pytest_cache/
+cover/
+
+# Translations
+*.mo
+*.pot
+
+# Django stuff:
+*.log
+local_settings.py
+db.sqlite3
+db.sqlite3-journal
+
+# Flask stuff:
+instance/
+.webassets-cache
+
+# Scrapy stuff:
+.scrapy
+
+# Sphinx documentation
+docs/_build/
+
+# PyBuilder
+.pybuilder/
+target/
+
+# Jupyter Notebook
+.ipynb_checkpoints
+
+# IPython
+profile_default/
+ipython_config.py
+
+# pyenv
+#   For a library or package, you might want to ignore these files since the code is
+#   intended to run in multiple environments; otherwise, check them in:
+# .python-version
+
+# pipenv
+#   According to pypa/pipenv#598, it is recommended to include Pipfile.lock in version control.
+#   However, in case of collaboration, if having platform-specific dependencies or dependencies
+#   having no cross-platform support, pipenv may install dependencies that don't work, or not
+#   install all needed dependencies.
+#Pipfile.lock
+
+# poetry
+#   Similar to Pipfile.lock, it is generally recommended to include poetry.lock in version control.
+#   This is especially recommended for binary packages to ensure reproducibility, and is more
+#   commonly ignored for libraries.
+#   https://python-poetry.org/docs/basic-usage/#commit-your-poetrylock-file-to-version-control
+#poetry.lock
+
+# pdm
+#   Similar to Pipfile.lock, it is generally recommended to include pdm.lock in version control.
+#pdm.lock
+#   pdm stores project-wide configurations in .pdm.toml, but it is recommended to not include it
+#   in version control.
+#   https://pdm.fming.dev/latest/usage/project/#working-with-version-control
+.pdm.toml
+.pdm-python
+.pdm-build/
+
+# PEP 582; used by e.g. github.com/David-OConnor/pyflow and github.com/pdm-project/pdm
+__pypackages__/
+
+# Celery stuff
+celerybeat-schedule
+celerybeat.pid
+
+# SageMath parsed files
+*.sage.py
+
+# Environments
+.env
+.venv
+env/
+venv/
+ENV/
+env.bak/
+venv.bak/
+
+# Spyder project settings
+.spyderproject
+.spyproject
+
+# Rope project settings
+.ropeproject
+
+# mkdocs documentation
+/site
+
+# mypy
+.mypy_cache/
+.dmypy.json
+dmypy.json
+
+# Pyre type checker
+.pyre/
+
+# pytype static type analyzer
+.pytype/
+
+# Cython debug symbols
+cython_debug/
+
+# PyCharm
+#  JetBrains specific template is maintained in a separate JetBrains.gitignore that can
+#  be found at https://github.com/github/gitignore/blob/main/Global/JetBrains.gitignore
+#  and can be added to the global gitignore or merged into this file.  For a more nuclear
+#  option (not recommended) you can uncomment the following to ignore the entire idea folder.
+#.idea/

.python-version

@@ -0,0 +1 @@
+3.10

README.md

@@ -0,0 +1,85 @@
+# mcp-server-qdrant: A Qdrant MCP server
+
+> The [Model Context Protocol (MCP)](https://modelcontextprotocol.io/introduction) is an open protocol that enables seamless integration between LLM applications and external data sources and tools. Whether you’re building an AI-powered IDE, enhancing a chat interface, or creating custom AI workflows, MCP provides a standardized way to connect LLMs with the context they need.
+
+This repository is an example of how to create a MCP server for [Qdrant](https://qdrant.tech/), a vector search engine.
+
+## Overview
+
+A basic Model Context Protocol server for keeping and retrieving memories in the Qdrant vector search engine.
+It acts as a semantic memory layer on top of the Qdrant database.
+
+## Components
+
+### Tools
+
+1. `qdrant-store-memory`
+   - Store a memory in the Qdrant database
+   - Input:
+     - `information` (string): Memory to store
+   - Returns: Confirmation message
+2. `qdrant-find-memories`
+   - Retrieve a memory from the Qdrant database
+   - Input:
+     - `query` (string): Query to retrieve a memory
+   - Returns: Memories stored in the Qdrant database as separate messages
+
+## Installation
+
+### Using uv (recommended)
+
+When using [`uv`](https://docs.astral.sh/uv/) no specific installation is needed to directly run *mcp-server-qdrant*.
+
+```shell
+uv run mcp-server-qdrant \
+  --qdrant-url "http://localhost:6333" \
+  --qdrant-api-key "your_api_key" \
+  --collection-name "my_collection" \
+  --fastembed-model-name "sentence-transformers/all-MiniLM-L6-v2"
+```
+
+## Usage with Claude Desktop
+
+To use this server with the Claude Desktop app, add the following configuration to the "mcpServers" section of your `claude_desktop_config.json`:
+
+```json
+{
+  "qdrant": {
+    "command": "uvx",
+    "args": [
+      "mcp-server-qdrant", 
+      "--qdrant-url", 
+      "http://localhost:6333",
+      "--qdrant-api-key", 
+      "your_api_key",
+      "--collection-name",
+      "your_collection_name"
+    ]
+  }
+}
+```
+
+Replace `http://localhost:6333`, `your_api_key` and `your_collection_name` with your Qdrant server URL, Qdrant API key 
+and collection name, respectively. The use of API key is optional, but recommended for security reasons, and depends on 
+the Qdrant server configuration.
+
+This MCP server will automatically create a collection with the specified name if it doesn't exist.
+
+By default, the server will use the `sentence-transformers/all-MiniLM-L6-v2` embedding model to encode memories.
+For the time being, only [FastEmbed](https://qdrant.github.io/fastembed/) models are supported, and you can change it
+by passing the `--fastembed-model-name` argument to the server.
+
+### Environment Variables
+
+The configuration of the server can be also done using environment variables:
+
+- `QDRANT_URL`: URL of the Qdrant server
+- `QDRANT_API_KEY`: API key for the Qdrant server
+- `COLLECTION_NAME`: Name of the collection to use
+- `FASTEMBED_MODEL_NAME`: Name of the FastEmbed model to use
+
+## License
+
+This MCP server is licensed under the MIT License. This means you are free to use, modify, and distribute the software, 
+subject to the terms and conditions of the MIT License. For more details, please see the LICENSE file in the project 
+repository.

pyproject.toml

@@ -0,0 +1,20 @@
+[project]
+name = "mcp-server-qdrant"
+version = "0.5.1"
+description = "MCP server for retrieving context from a Qdrant vector database"
+readme = "README.md"
+requires-python = ">=3.10"
+dependencies = [
+    "mcp>=0.9.1",
+    "qdrant-client[fastembed]>=1.12.0",
+]
+
+[build-system]
+requires = ["hatchling"]
+build-backend = "hatchling.build"
+
+[tool.uv]
+dev-dependencies = ["pyright>=1.1.389", "pytest>=8.3.3", "ruff>=0.8.0"]
+
+[project.scripts]
+mcp-server-qdrant = "mcp_server_qdrant:main"

src/mcp_server_qdrant/__init__.py

@@ -0,0 +1,10 @@
+from . import server
+
+
+def main():
+    """Main entry point for the package."""
+    server.main()
+
+
+# Optionally expose other important items at package level
+__all__ = ["main", "server"]

src/mcp_server_qdrant/qdrant.py

@@ -0,0 +1,52 @@
+from typing import Optional
+from qdrant_client import AsyncQdrantClient, models
+
+
+class QdrantConnector:
+    """
+    Encapsulates the connection to a Qdrant server and all the methods to interact with it.
+    :param qdrant_url: The URL of the Qdrant server.
+    :param qdrant_api_key: The API key to use for the Qdrant server.
+    :param collection_name: The name of the collection to use.
+    :param fastembed_model_name: The name of the FastEmbed model to use.
+    """
+
+    def __init__(
+        self,
+        qdrant_url: str,
+        qdrant_api_key: Optional[str],
+        collection_name: str,
+        fastembed_model_name: str,
+    ):
+        self._qdrant_url = qdrant_url.rstrip("/")
+        self._qdrant_api_key = qdrant_api_key
+        self._collection_name = collection_name
+        self._fastembed_model_name = fastembed_model_name
+        # For the time being, FastEmbed models are the only supported ones.
+        # A list of all available models can be found here:
+        # https://qdrant.github.io/fastembed/examples/Supported_Models/
+        self._client = AsyncQdrantClient(qdrant_url, api_key=qdrant_api_key)
+        self._client.set_model(fastembed_model_name)
+
+    async def store_memory(self, information: str):
+        """
+        Store a memory in the Qdrant collection.
+        :param information: The information to store.
+        """
+        await self._client.add(
+            self._collection_name,
+            documents=[information],
+        )
+
+    async def find_memories(self, query: str) -> list[str]:
+        """
+        Find memories in the Qdrant collection.
+        :param query: The query to use for the search.
+        :return: A list of memories found.
+        """
+        search_results = await self._client.query(
+            self._collection_name,
+            query_text=query,
+            limit=10,
+        )
+        return [result.document for result in search_results]

src/mcp_server_qdrant/server.py

@@ -0,0 +1,164 @@
+from typing import Optional
+
+from mcp.server import Server, NotificationOptions
+from mcp.server.models import InitializationOptions
+
+import click
+import mcp.types as types
+import asyncio
+import mcp
+
+from .qdrant import QdrantConnector
+
+
+def serve(
+    qdrant_url: str,
+    qdrant_api_key: Optional[str],
+    collection_name: str,
+    fastembed_model_name: str,
+) -> Server:
+    """
+    Instantiate the server and configure tools to store and find memories in Qdrant.
+    :param qdrant_url: The URL of the Qdrant server.
+    :param qdrant_api_key: The API key to use for the Qdrant server.
+    :param collection_name: The name of the collection to use.
+    :param fastembed_model_name: The name of the FastEmbed model to use.
+    """
+    server = Server("qdrant")
+
+    qdrant = QdrantConnector(
+        qdrant_url, qdrant_api_key, collection_name, fastembed_model_name
+    )
+
+    @server.list_tools()
+    async def handle_list_tools() -> list[types.Tool]:
+        """
+        Return the list of tools that the server provides. By default, there are two
+        tools: one to store memories and another to find them. Finding the memories is not
+        implemented as a resource, as it requires a query to be passed and resources point
+        to a very specific piece of data.
+        """
+        return [
+            types.Tool(
+                name="qdrant-store-memory",
+                description=(
+                    "Keep the memory for later use, when you are asked to remember something."
+                ),
+                inputSchema={
+                    "type": "object",
+                    "properties": {
+                        "information": {
+                            "type": "string",
+                        },
+                    },
+                    "required": ["information"],
+                },
+            ),
+            types.Tool(
+                name="qdrant-find-memories",
+                description=(
+                    "Look up memories in Qdrant. Use this tool when you need to: \n"
+                    " - Find memories by their content \n"
+                    " - Access memories for further analysis \n"
+                    " - Get some personal information about the user"
+                ),
+                inputSchema={
+                    "type": "object",
+                    "properties": {
+                        "query": {
+                            "type": "string",
+                            "description": "The query to search for in the memories",
+                        },
+                    },
+                    "required": ["query"],
+                },
+            ),
+        ]
+
+    @server.call_tool()
+    async def handle_tool_call(
+        name: str, arguments: dict | None
+    ) -> list[types.TextContent | types.ImageContent | types.EmbeddedResource]:
+        if name not in ["qdrant-store-memory", "qdrant-find-memories"]:
+            raise ValueError(f"Unknown tool: {name}")
+
+        if name == "qdrant-store-memory":
+            if not arguments or "information" not in arguments:
+                raise ValueError("Missing required argument 'information'")
+            information = arguments["information"]
+            await qdrant.store_memory(information)
+            return [types.TextContent(type="text", text=f"Remembered: {information}")]
+
+        if name == "qdrant-find-memories":
+            if not arguments or "query" not in arguments:
+                raise ValueError("Missing required argument 'query'")
+            query = arguments["query"]
+            memories = await qdrant.find_memories(query)
+            content = [
+                types.TextContent(
+                    type="text", text=f"Memories for the query '{query}'"
+                ),
+            ]
+            for memory in memories:
+                content.append(
+                    types.TextContent(type="text", text=f"<memory>{memory}</memory>")
+                )
+            return content
+
+    return server
+
+
+@click.command()
+@click.option(
+    "--qdrant-url",
+    envvar="QDRANT_URL",
+    required=True,
+    help="Qdrant URL",
+)
+@click.option(
+    "--qdrant-api-key",
+    envvar="QDRANT_API_KEY",
+    required=False,
+    help="Qdrant API key",
+)
+@click.option(
+    "--collection-name",
+    envvar="COLLECTION_NAME",
+    required=True,
+    help="Collection name",
+)
+@click.option(
+    "--fastembed-model-name",
+    envvar="FASTEMBED_MODEL_NAME",
+    required=True,
+    help="FastEmbed model name",
+    default="sentence-transformers/all-MiniLM-L6-v2",
+)
+def main(
+    qdrant_url: str,
+    qdrant_api_key: str,
+    collection_name: Optional[str],
+    fastembed_model_name: str,
+):
+    async def _run():
+        async with mcp.server.stdio.stdio_server() as (read_stream, write_stream):
+            server = serve(
+                qdrant_url,
+                qdrant_api_key,
+                collection_name,
+                fastembed_model_name,
+            )
+            await server.run(
+                read_stream,
+                write_stream,
+                InitializationOptions(
+                    server_name="qdrant",
+                    server_version="0.5.1",
+                    capabilities=server.get_capabilities(
+                        notification_options=NotificationOptions(),
+                        experimental_capabilities={},
+                    ),
+                ),
+            )
+
+    asyncio.run(_run())