### Create a New Collection

Source: https://context7.com/zilliztech/zilliz-mcp-server/llms.txt

Create a new collection with specified schema using Quick Setup. Supports configurable dimensions, metric types, and field names. This example demonstrates a basic collection with COSINE similarity.

```python
# MCP Tool Call - Basic collection with COSINE similarity
result = await create_collection(
    cluster_id="inxx-xxxxxxxxxxxxxxx",
    region_id="gcp-us-west1",
    endpoint="https://inxx-xxxxxxxxxxxxxxx.api.gcp-us-west1.zillizcloud.com",
    collection_name="product_embeddings",
    dimension=768,  # Vector dimension (e.g., 768 for BERT embeddings)
    metric_type="COSINE",  # Options: L2, IP, COSINE
    id_type="Int64",  # Options: Int64, VarChar
    auto_id=True,
    primary_field_name="id",
    vector_field_name="vector"
)
```

--------------------------------

### Verify uv Installation

Source: https://github.com/zilliztech/zilliz-mcp-server/blob/master/docs/USERGUIDE.md

Confirm that the 'uv' command-line tool is installed and accessible in your system's PATH.

```bash
uv --version
```

--------------------------------

### create_collection

Source: https://context7.com/zilliztech/zilliz-mcp-server/llms.txt

Create a new collection with specified schema using Quick Setup.

```APIDOC
## create_collection

### Description
Create a new collection with specified schema using Quick Setup. Supports configurable dimensions, metric types, and field names.

### Parameters
#### Request Body
- **cluster_id** (string) - Required
- **collection_name** (string) - Required
- **dimension** (integer) - Required
- **metric_type** (string) - Required
- **id_type** (string) - Required
- **auto_id** (boolean) - Required
- **primary_field_name** (string) - Required
- **vector_field_name** (string) - Required
```

--------------------------------

### Clone and Setup Zilliz MCP Server

Source: https://github.com/zilliztech/zilliz-mcp-server/blob/master/README.md

Commands to clone the repository and prepare the environment configuration.

```bash
git clone https://github.com/zilliztech/zilliz-mcp-server.git
cd zilliz-mcp-server
```

```bash
cp example.env .env
```

--------------------------------

### Clone and Run HTTP Server

Source: https://context7.com/zilliztech/zilliz-mcp-server/llms.txt

Clone the repository, set up environment variables, and start the Zilliz MCP Server as a standalone HTTP service.

```bash
# Clone the repository
git clone https://github.com/zilliztech/zilliz-mcp-server.git
cd zilliz-mcp-server

# Create environment file
cp example.env .env
# Edit .env and add: ZILLIZ_API_KEY="your_api_key_here"

# Start HTTP server
uv run src/zilliz_mcp_server/server.py --transport streamable-http
```

--------------------------------

### Start MCP Server as HTTP Service

Source: https://github.com/zilliztech/zilliz-mcp-server/blob/master/README.md

Execute the server using the streamable-http transport mode.

```bash
uv run src/zilliz_mcp_server/server.py --transport streamable-http
```

--------------------------------

### Verify Python Version

Source: https://github.com/zilliztech/zilliz-mcp-server/blob/master/docs/USERGUIDE.md

Check if your Python installation meets the minimum version requirement of 3.10.

```bash
python --version
```

--------------------------------

### Mock External Dependencies in Tests

Source: https://github.com/zilliztech/zilliz-mcp-server/blob/master/TESTING.md

Example of using patch to mock the openapi_client for testing tool functionality.

```python
@patch('zilliz_mcp_server.tools.zilliz.zilliz_tools.openapi_client')
async def test_list_projects_success(self, mock_client, sample_project_response):
    mock_client.control_plane_api_request.return_value = sample_project_response
    # Test implementation...
```

--------------------------------

### Query Entities with Pagination

Source: https://context7.com/zilliztech/zilliz-mcp-server/llms.txt

Perform queries with pagination to retrieve entities in batches. Specify limit and offset for controlling the number of results and their starting point.

```python
result = await query(
    cluster_id="inxx-xxxxxxxxxxxxxxx",
    region_id="gcp-us-west1",
    endpoint="https://inxx-xxxxxxxxxxxxxxx.api.gcp-us-west1.zillizcloud.com",
    collection_name="product_embeddings",
    filter="id in [1, 2, 3, 4, 5]",
    output_fields=["product_name", "vector"],
    limit=10,
    offset=0
)
```

--------------------------------

### Configure API Key

Source: https://github.com/zilliztech/zilliz-mcp-server/blob/master/README.md

Add the Zilliz Cloud API key to the .env file.

```text
ZILLIZ_API_KEY="your_api_key_here"
```

--------------------------------

### List Projects using MCP

Source: https://context7.com/zilliztech/zilliz-mcp-server/llms.txt

Call the `list_projects` tool to retrieve a list of all projects associated with your Zilliz Cloud API Key. The response includes project details like name, ID, instance count, and creation time.

```python
# MCP Tool Call
result = await list_projects()

# Expected Response
[
    {
        "project_name": "Default Project",
        "project_id": "proj-f5b02814db7ccfe2d16293",
        "instance_count": 2,
        "create_time": "2023-06-14T06:59:07Z"
    },
    {
        "project_name": "Production",
        "project_id": "proj-a1b2c3d4e5f6g7h8i9j0",
        "instance_count": 5,
        "create_time": "2024-01-15T10:30:00Z"
    }
]
```

--------------------------------

### Test ZillizConfig with Valid Environment Variables

Source: https://github.com/zilliztech/zilliz-mcp-server/blob/master/TESTING.md

Tests the initialization of ZillizConfig using valid environment variables. Ensure that the configuration is loaded correctly.

```python
def test_init_with_valid_env_vars(self, mock_env_vars):
    with patch.dict(os.environ, mock_env_vars, clear=True):
        config = ZillizConfig()
        assert config.token == "test-token-123"
```

--------------------------------

### MCP Client Configuration

Source: https://context7.com/zilliztech/zilliz-mcp-server/llms.txt

Configure your MCP client to connect to the Zilliz MCP Server. Ensure your ZILLIZ_CLOUD_TOKEN is set in the environment.

```json
{
  "mcpServers": {
    "zilliz-mcp-server": {
      "command": "uvx",
      "args": ["zilliz-mcp-server"],
      "env": {
        "ZILLIZ_CLOUD_TOKEN": "your-api-token-here"
      }
    }
  }
}
```

--------------------------------

### Enable Asyncio for Tests

Source: https://github.com/zilliztech/zilliz-mcp-server/blob/master/TESTING.md

Configuration marker to enable asyncio support for all tests within a module.

```python
# Enable asyncio for all async tests in a module
pytestmark = pytest.mark.asyncio
```

--------------------------------

### Clone Zilliz MCP Server Project

Source: https://github.com/zilliztech/zilliz-mcp-server/blob/master/docs/USERGUIDE.md

Clone the Zilliz MCP Server repository and navigate into the project directory.

```bash
git clone https://github.com/zilliztech/zilliz-mcp-server.git
cd zilliz-mcp-server
```

--------------------------------

### Execute Test Commands

Source: https://github.com/zilliztech/zilliz-mcp-server/blob/master/TESTING.md

Commands for running the full test suite or specific test categories using uv.

```bash
# Run all tests
./run_tests.sh

# Or manually:
uv run pytest tests/unit/ -v
```

```bash
# Configuration tests only
uv run pytest tests/unit/test_settings.py -v

# HTTP client tests only
uv run pytest tests/unit/test_openapi_client.py -v

# API tools tests only
uv run pytest tests/unit/tools/ -v

# Run tests with coverage
uv add pytest-cov
uv run pytest tests/unit/ --cov=src/zilliz_mcp_server --cov-report=html
```

--------------------------------

### Configure MCP Client Connection

Source: https://github.com/zilliztech/zilliz-mcp-server/blob/master/README.md

JSON configuration for connecting an MCP client to the running server.

```json
{
  "mcpServers": {
    "zilliz-mcp-server": {
      "url": "http://localhost:8000/mcp",
      "transport": "streamable-http",
      "description": "Zilliz Cloud and Milvus MCP Server"
    }
  }
}
```

--------------------------------

### Create Free Tier Cluster using MCP

Source: https://context7.com/zilliztech/zilliz-mcp-server/llms.txt

Use the `create_free_cluster` tool to provision a new free-tier Milvus cluster. Requires a cluster name and project ID. The region is set by the `ZILLIZ_CLOUD_FREE_CLUSTER_REGION` environment variable.

```python
# MCP Tool Call
result = await create_free_cluster(
    cluster_name="my-test-cluster",
    project_id="proj-f5b02814db7ccfe2d16293"
)

# Expected Response
{
    "cluster_id": "inxx-abc123def456",
    "username": "db_admin_user",
    "prompt": "successfully submitted, cluster is being created..."
}
```

--------------------------------

### Configure mcp.json for Cursor

Source: https://github.com/zilliztech/zilliz-mcp-server/blob/master/docs/USERGUIDE.md

Add the Zilliz MCP Server configuration to your Cursor editor's mcp.json file. Ensure you replace 'your-token-here' with your actual Zilliz API key.

```json
{
  "mcpServers": {
    "zilliz-mcp-server": {
      "command": "uvx",
      "args": ["zilliz-mcp-server"],
      "env": {
          "ZILLIZ_CLOUD_TOKEN": "your-token-here"
      }
    }
  }
}
```

--------------------------------

### Create Collection with VarChar Primary Key

Source: https://context7.com/zilliztech/zilliz-mcp-server/llms.txt

Use this to create a new collection with a VarChar primary key. Ensure the dimension matches your embedding model.

```python
result = await create_collection(
    cluster_id="inxx-xxxxxxxxxxxxxxx",
    region_id="gcp-us-west1",
    endpoint="https://inxx-xxxxxxxxxxxxxxx.api.gcp-us-west1.zillizcloud.com",
    collection_name="document_store",
    dimension=1536,  # OpenAI embedding dimension
    metric_type="IP",  # Inner Product
    id_type="VarChar",
    auto_id=False,
    primary_field_name="doc_id",
    vector_field_name="embedding"
)
```

--------------------------------

### Environment Variables for Zilliz MCP Server

Source: https://context7.com/zilliztech/zilliz-mcp-server/llms.txt

Configure Zilliz MCP Server behavior using environment variables. Required variables include ZILLIZ_CLOUD_TOKEN for authentication.

```bash
# Required: Zilliz Cloud API Token
ZILLIZ_CLOUD_TOKEN="your-api-token"

# Optional: Zilliz Cloud API endpoint (default: https://api.cloud.zilliz.com)
ZILLIZ_CLOUD_URI="https://api.cloud.zilliz.com"

# Optional: Default region for free cluster creation (default: gcp-us-west1)
ZILLIZ_CLOUD_FREE_CLUSTER_REGION="gcp-us-west1"

# Optional: MCP server configuration
MCP_SERVER_HOST="localhost"
MCP_SERVER_PORT=8000
```

--------------------------------

### List Clusters with Pagination using MCP

Source: https://context7.com/zilliztech/zilliz-mcp-server/llms.txt

Call the `list_clusters` tool with pagination parameters to retrieve cluster details. This includes cluster ID, name, status, connection address, and plan type.

```python
# MCP Tool Call with pagination
result = await list_clusters(page_size=10, current_page=1)

# Expected Response
[
    {
        "cluster_id": "inxx-xxxxxxxxxxxxxxx",
        "cluster_name": "dedicated-3",
        "description": "Production vector search cluster",
        "region_id": "aws-us-west-2",
        "plan": "Standard",
        "cu_type": "Performance-optimized",
        "cu_size": 1,
        "status": "RUNNING",
        "connect_address": "https://inxx-xxxxxxxxxxxxxxx.aws-us-west-2.vectordb.zillizcloud.com:19530",
        "private_link_address": "",
        "project_id": "proj-xxxxxxxxxxxxxxxxxxxxxx",
        "create_time": "2024-06-30T16:49:50Z"
    }
]
```

--------------------------------

### Basic Vector Search with MCP Tool

Source: https://context7.com/zilliztech/zilliz-mcp-server/llms.txt

Perform a basic vector search to find similar items. Specify cluster, region, endpoint, collection, query data, and desired output fields.

```python
result = await search(
    cluster_id="inxx-xxxxxxxxxxxxxxx",
    region_id="gcp-us-west1",
    endpoint="https://inxx-xxxxxxxxxxxxxxx.api.gcp-us-west1.zillizcloud.com",
    collection_name="product_embeddings",
    data=[[0.1, 0.2, 0.3, ...]],  # Query vector(s)
    anns_field="vector",
    limit=10,
    output_fields=["product_name", "category", "price"]
)
```

--------------------------------

### create_free_cluster

Source: https://context7.com/zilliztech/zilliz-mcp-server/llms.txt

Creates a new free-tier Milvus cluster in Zilliz Cloud.

```APIDOC
## create_free_cluster

### Description
Create a new free-tier Milvus cluster in Zilliz Cloud. Requires a cluster name and project ID.

### Parameters
#### Request Body
- **cluster_name** (string) - Required - Name for the new cluster
- **project_id** (string) - Required - ID of the project to host the cluster

### Response
#### Success Response (200)
- **cluster_id** (string) - ID of the created cluster
- **username** (string) - Admin username for the cluster
- **prompt** (string) - Status message

### Response Example
{
    "cluster_id": "inxx-abc123def456",
    "username": "db_admin_user",
    "prompt": "successfully submitted, cluster is being created..."
}
```

--------------------------------

### list_projects

Source: https://context7.com/zilliztech/zilliz-mcp-server/llms.txt

Retrieves a list of all projects associated with the provided API Key.

```APIDOC
## list_projects

### Description
List all projects scoped to your API Key in Zilliz Cloud. Returns project names, IDs, instance counts, and creation timestamps.

### Response
#### Success Response (200)
- **project_name** (string) - Name of the project
- **project_id** (string) - Unique identifier for the project
- **instance_count** (integer) - Number of instances in the project
- **create_time** (string) - ISO 8601 timestamp of creation

### Response Example
[
    {
        "project_name": "Default Project",
        "project_id": "proj-f5b02814db7ccfe2d16293",
        "instance_count": 2,
        "create_time": "2023-06-14T06:59:07Z"
    }
]
```

--------------------------------

### Vector Search with Pagination and Grouping

Source: https://context7.com/zilliztech/zilliz-mcp-server/llms.txt

Implement pagination and grouping in vector search to manage large result sets and categorize findings. Use offset for pagination and grouping_field for categorization.

```python
result = await search(
    cluster_id="inxx-xxxxxxxxxxxxxxx",
    region_id="gcp-us-west1",
    endpoint="https://inxx-xxxxxxxxxxxxxxx.api.gcp-us-west1.zillizcloud.com",
    collection_name="product_embeddings",
    data=[[0.1, 0.2, 0.3, ...]],
    anns_field="vector",
    limit=10,
    offset=20,
    grouping_field="category",
    output_fields=["product_name", "category"],
    consistency_level="Strong"
)
```

--------------------------------

### Create Collection with VarChar Primary Key

Source: https://context7.com/zilliztech/zilliz-mcp-server/llms.txt

This endpoint allows the creation of a new collection with a specified VarChar primary key. It configures collection properties such as dimension, metric type, and field names for IDs and vectors.

```APIDOC
## POST /api/collections/create

### Description
Creates a new collection with a VarChar primary key.

### Method
POST

### Endpoint
/api/collections/create

### Parameters
#### Request Body
- **cluster_id** (string) - Required - The ID of the Zilliz cluster.
- **region_id** (string) - Required - The region where the cluster is located.
- **endpoint** (string) - Required - The API endpoint for the Zilliz cluster.
- **collection_name** (string) - Required - The name of the collection to create.
- **dimension** (integer) - Required - The dimension of the vector embeddings.
- **metric_type** (string) - Required - The similarity metric type (e.g., IP, L2, COSINE).
- **id_type** (string) - Required - The data type of the primary key (e.g., VarChar, Int64).
- **auto_id** (boolean) - Required - Whether to automatically generate IDs.
- **primary_field_name** (string) - Required - The name of the primary key field.
- **vector_field_name** (string) - Required - The name of the vector field.

### Request Example
```json
{
  "cluster_id": "inxx-xxxxxxxxxxxxxxx",
  "region_id": "gcp-us-west1",
  "endpoint": "https://inxx-xxxxxxxxxxxxxxx.api.gcp-us-west1.zillizcloud.com",
  "collection_name": "document_store",
  "dimension": 1536,
  "metric_type": "IP",
  "id_type": "VarChar",
  "auto_id": false,
  "primary_field_name": "doc_id",
  "vector_field_name": "embedding"
}
```

### Response
#### Success Response (200)
- **code** (integer) - The status code of the operation.
- **data** (object) - An empty object upon successful creation.

#### Response Example
```json
{
  "code": 0,
  "data": {}
}
```
```

--------------------------------

### Describe Collection Details

Source: https://context7.com/zilliztech/zilliz-mcp-server/llms.txt

Retrieve detailed information about an existing collection, including its schema, fields, and load status. The `db_name` is optional for free/serverless clusters.

```python
result = await describe_collection(
    cluster_id="inxx-xxxxxxxxxxxxxxx",
    region_id="gcp-us-west1",
    endpoint="https://inxx-xxxxxxxxxxxxxxx.api.gcp-us-west1.zillizcloud.com",
    collection_name="product_embeddings",
    db_name=""  # Optional for free/serverless clusters
)
```

--------------------------------

### Project Test Directory Structure

Source: https://github.com/zilliztech/zilliz-mcp-server/blob/master/TESTING.md

Visual representation of the test suite organization within the repository.

```text
tests/
├── __init__.py
├── conftest.py              # pytest configuration and shared fixtures
├── pytest.ini              # pytest settings
├── unit/
│   ├── __init__.py
│   ├── test_settings.py     # Configuration validation tests (12 tests)
│   ├── test_openapi_client.py  # HTTP client tests (23 tests)
│   └── tools/
│       ├── __init__.py
│       ├── test_zilliz_tools.py    # Zilliz Cloud API tests (12 tests)
│       └── test_milvus_tools.py    # Milvus data operations tests (15 tests)
```

--------------------------------

### resume_cluster

Source: https://context7.com/zilliztech/zilliz-mcp-server/llms.txt

Resume a suspended dedicated cluster. The cluster will take several minutes to become available.

```APIDOC
## resume_cluster

### Description
Resume a suspended dedicated cluster. The cluster will take several minutes to become available.

### Parameters
#### Request Body
- **cluster_id** (string) - Required - The ID of the cluster to resume.

### Request Example
{
  "cluster_id": "inxx-xxxxxxxxxxxxxxx"
}

### Response
#### Success Response (200)
- **cluster_id** (string) - The ID of the cluster.
- **prompt** (string) - Status update message.

#### Response Example
{
  "cluster_id": "inxx-xxxxxxxxxxxxxxx",
  "prompt": "successfully Submitted. Cluster is being resumed, which is expected to takes several minutes."
}
```

--------------------------------

### Describe Cluster using MCP

Source: https://context7.com/zilliztech/zilliz-mcp-server/llms.txt

Call the `describe_cluster` tool with a specific `cluster_id` to retrieve detailed information about a Milvus cluster, including its status, connection details, and storage information.

```python
# MCP Tool Call
result = await describe_cluster(cluster_id="inxx-xxxxxxxxxxxxxxx")
```

--------------------------------

### List Databases in a Cluster

Source: https://context7.com/zilliztech/zilliz-mcp-server/llms.txt

List all databases within a specific cluster. This function returns an array of database names.

```python
# MCP Tool Call
result = await list_databases(
    cluster_id="inxx-xxxxxxxxxxxxxxx",
    region_id="gcp-us-west1",
    endpoint="https://inxx-xxxxxxxxxxxxxxx.api.gcp-us-west1.zillizcloud.com"
)
```

--------------------------------

### list_databases

Source: https://context7.com/zilliztech/zilliz-mcp-server/llms.txt

List all databases within a specific cluster.

```APIDOC
## list_databases

### Description
List all databases within a specific cluster. Returns an array of database names.

### Parameters
#### Request Body
- **cluster_id** (string) - Required
- **region_id** (string) - Required
- **endpoint** (string) - Required

### Response
#### Success Response (200)
- **databases** (array) - List of database names.
```

--------------------------------

### Zilliz Control Plane Tools

Source: https://github.com/zilliztech/zilliz-mcp-server/blob/master/README.md

Tools for managing Zilliz Cloud resources such as projects and clusters.

```APIDOC
## Zilliz Control Plane Tools

### Description
These tools allow AI agents to manage Zilliz Cloud infrastructure.

### Available Tools
- **list_projects**: List all projects in your Zilliz Cloud account.
- **list_clusters**: List all clusters within your projects.
- **create_free_cluster**: Create a new, free-tier Milvus cluster.
- **describe_cluster**: Get detailed information about a specific cluster.
- **suspend_cluster**: Suspend a running cluster to save costs.
- **resume_cluster**: Resume a suspended cluster.
- **query_cluster_metrics**: Query various performance metrics for a cluster.
```

--------------------------------

### Test ZillizConfig with Invalid Port Range

Source: https://github.com/zilliztech/zilliz-mcp-server/blob/master/TESTING.md

Tests the exception handling for an invalid port range when initializing ZillizConfig. This ensures that the port number is validated correctly.

```python
def test_invalid_port_range(self):
    env_vars = {"ZILLIZ_CLOUD_TOKEN": "test-token", "MCP_SERVER_PORT": "99999"}
    with patch.dict(os.environ, env_vars, clear=True):
        with pytest.raises(ValueError, match="MCP_SERVER_PORT must be between 1 and 65535"):
            ZillizConfig()
```

--------------------------------

### List Collections in a Database

Source: https://context7.com/zilliztech/zilliz-mcp-server/llms.txt

List all collection names within a specified database. This function returns an array of collection names. The `db_name` parameter is optional for free/serverless clusters.

```python
# MCP Tool Call
result = await list_collections(
    cluster_id="inxx-xxxxxxxxxxxxxxx",
    region_id="gcp-us-west1",
    endpoint="https://inxx-xxxxxxxxxxxxxxx.api.gcp-us-west1.zillizcloud.com",
    db_name="default"  # Optional for free/serverless clusters
)
```

--------------------------------

### list_clusters

Source: https://context7.com/zilliztech/zilliz-mcp-server/llms.txt

Lists all clusters scoped to the API Key with support for pagination.

```APIDOC
## list_clusters

### Description
List all clusters scoped to your API Key with pagination support. Returns cluster details including status, connection address, plan type, and compute unit information.

### Parameters
#### Query Parameters
- **page_size** (integer) - Optional - Number of items per page
- **current_page** (integer) - Optional - Current page number

### Response
#### Success Response (200)
- **cluster_id** (string) - Unique identifier for the cluster
- **cluster_name** (string) - Name of the cluster
- **status** (string) - Current status of the cluster
- **connect_address** (string) - Connection URL for the cluster

### Response Example
[
    {
        "cluster_id": "inxx-xxxxxxxxxxxxxxx",
        "cluster_name": "dedicated-3",
        "status": "RUNNING",
        "connect_address": "https://inxx-xxxxxxxxxxxxxxx.aws-us-west-2.vectordb.zillizcloud.com:19530"
    }
]
```

--------------------------------

### Milvus Data Plane Tools

Source: https://github.com/zilliztech/zilliz-mcp-server/blob/master/README.md

Tools for interacting with data stored within a Milvus cluster.

```APIDOC
## Milvus Data Plane Tools

### Description
These tools enable data operations within Milvus collections.

### Available Tools
- **list_databases**: List all databases within a specific cluster.
- **list_collections**: List all collections within a database.
- **create_collection**: Create a new collection with a specified schema.
- **describe_collection**: Get detailed information about a collection, including its schema.
- **insert_entities**: Insert entities (data records with vectors) into a collection.
- **delete_entities**: Delete entities from a collection based on IDs or a filter expression.
- **search**: Perform a vector similarity search on a collection.
- **query**: Query entities based on a scalar filter expression.
- **hybrid_search**: Perform a hybrid search combining vector similarity and scalar filters.
```

--------------------------------

### Resume a Suspended Cluster

Source: https://context7.com/zilliztech/zilliz-mcp-server/llms.txt

Use this function to resume a suspended dedicated cluster. The cluster may take several minutes to become fully available.

```python
# MCP Tool Call
result = await resume_cluster(cluster_id="inxx-xxxxxxxxxxxxxxx")
```

--------------------------------

### list_collections

Source: https://context7.com/zilliztech/zilliz-mcp-server/llms.txt

List all collection names within a database.

```APIDOC
## list_collections

### Description
List all collection names within a database. Returns an array of collection names for the specified database.

### Parameters
#### Request Body
- **cluster_id** (string) - Required
- **region_id** (string) - Required
- **endpoint** (string) - Required
- **db_name** (string) - Optional

### Response
#### Success Response (200)
- **collections** (array) - List of collection names.
```

--------------------------------

### Describe Collection

Source: https://context7.com/zilliztech/zilliz-mcp-server/llms.txt

Retrieves detailed information about a specified collection, including its schema, fields, indexes, and current load state.

```APIDOC
## GET /api/collections/describe

### Description
Gets detailed information about a collection including its schema, fields, indexes, and load state.

### Method
GET

### Endpoint
/api/collections/describe

### Parameters
#### Query Parameters
- **cluster_id** (string) - Required - The ID of the Zilliz cluster.
- **region_id** (string) - Required - The region where the cluster is located.
- **endpoint** (string) - Required - The API endpoint for the Zilliz cluster.
- **collection_name** (string) - Required - The name of the collection to describe.
- **db_name** (string) - Optional - The name of the database (if applicable).

### Response
#### Success Response (200)
- **code** (integer) - The status code of the operation.
- **data** (object) - An object containing detailed information about the collection.
  - **aliases** (array) - Aliases associated with the collection.
  - **autoId** (boolean) - Indicates if auto-ID is enabled.
  - **collectionID** (integer) - The unique identifier for the collection.
  - **collectionName** (string) - The name of the collection.
  - **consistencyLevel** (string) - The consistency level of the collection.
  - **description** (string) - A description of the collection.
  - **enableDynamicField** (boolean) - Indicates if dynamic fields are enabled.
  - **fields** (array) - An array of field objects, each describing a field in the collection.
  - **indexes** (array) - An array of index objects, describing the indexes on the collection.
  - **load** (string) - The current load state of the collection (e.g., LoadStateLoaded).
  - **partitionsNum** (integer) - The number of partitions in the collection.
  - **properties** (array) - Properties associated with the collection.

#### Response Example
```json
{
  "code": 0,
  "data": {
    "aliases": [],
    "autoId": true,
    "collectionID": 448707763883002000,
    "collectionName": "product_embeddings",
    "consistencyLevel": "Bounded",
    "description": "",
    "enableDynamicField": true,
    "fields": [
      {
        "autoId": true,
        "description": "",
        "id": 100,
        "name": "id",
        "partitionKey": false,
        "primaryKey": true,
        "type": "Int64"
      },
      {
        "autoId": false,
        "description": "",
        "id": 101,
        "name": "vector",
        "params": [{"key": "dim", "value": "768"}],
        "partitionKey": false,
        "primaryKey": false,
        "type": "FloatVector"
      }
    ],
    "indexes": [
      {
        "fieldName": "vector",
        "indexName": "vector",
        "metricType": "COSINE"
      }
    ],
    "load": "LoadStateLoaded",
    "partitionsNum": 1,
    "properties": []
  }
}
```
```

--------------------------------

### Query Cluster Performance Metrics

Source: https://context7.com/zilliztech/zilliz-mcp-server/llms.txt

Query performance metrics for a specific cluster. You can specify a time period or a specific time range, along with desired metrics and statistics.

```python
# MCP Tool Call - Query metrics for the last hour
result = await query_cluster_metrics(
    cluster_id="inxx-xxxxxxxxxxxxxxx",
    period="PT1H",  # ISO 8601 duration: 1 hour
    granularity="PT30S",  # 30-second intervals
    metric_queries=[
        {"metricName": "CU_COMPUTATION", "stat": "AVG"},
        {"metricName": "REQ_SEARCH_COUNT", "stat": "AVG"},
        {"metricName": "REQ_SEARCH_LATENCY_P99", "stat": "P99"},
        {"metricName": "STORAGE_USE", "stat": "AVG"}
    ]
)
```

```python
# Alternative: Query with specific time range
result = await query_cluster_metrics(
    cluster_id="inxx-xxxxxxxxxxxxxxx",
    start="2024-06-30T15:00:00Z",
    end="2024-06-30T16:00:00Z",
    granularity="PT1M",
    metric_queries=[
        {"metricName": "ENTITIES_COUNT", "stat": "AVG"},
        {"metricName": "COLLECTIONS_COUNT", "stat": "AVG"}
    ]
)
```

--------------------------------

### describe_cluster

Source: https://context7.com/zilliztech/zilliz-mcp-server/llms.txt

Retrieves detailed information about a specific cluster.

```APIDOC
## describe_cluster

### Description
Get detailed information about a specific cluster including status, connection address, storage size, and creation progress.

### Parameters
#### Request Body
- **cluster_id** (string) - Required - The ID of the cluster to describe
```

--------------------------------

### Hybrid Search with RRF Reranking

Source: https://context7.com/zilliztech/zilliz-mcp-server/llms.txt

Execute a hybrid search combining multiple vector similarity searches and rerank results using Reciprocal Rank Fusion (RRF). Configure search requests and RRF parameters.

```python
result = await hybrid_search(
    cluster_id="inxx-xxxxxxxxxxxxxxx",
    region_id="gcp-us-west1",
    endpoint="https://inxx-xxxxxxxxxxxxxxx.api.gcp-us-west1.zillizcloud.com",
    collection_name="multi_vector_collection",
    search_requests=[
        {
            "data": [[0.1, 0.2, 0.3, ...]],  # Text embedding
            "annsField": "text_vector",
            "limit": 20,
            "metricType": "COSINE"
        },
        {
            "data": [[0.4, 0.5, 0.6, ...]],  # Image embedding
            "annsField": "image_vector",
            "limit": 20,
            "metricType": "COSINE"
        }
    ],
    rerank_strategy="rrf",
    rerank_params={"k": 60},  # RRF constant
    limit=10,
    output_fields=["title", "description", "image_url"]
)
```

--------------------------------

### Hybrid Search with Weighted Reranking

Source: https://context7.com/zilliztech/zilliz-mcp-server/llms.txt

Perform a hybrid search with weighted reranking, assigning specific weights to different vector types (e.g., dense and sparse). Supports filtering within search requests.

```python
result = await hybrid_search(
    cluster_id="inxx-xxxxxxxxxxxxxxx",
    region_id="gcp-us-west1",
    endpoint="https://inxx-xxxxxxxxxxxxxxx.api.gcp-us-west1.zillizcloud.com",
    collection_name="multi_vector_collection",
    search_requests=[
        {
            "data": [[0.1, 0.2, 0.3, ...]],
            "annsField": "dense_vector",
            "limit": 50,
            "filter": "category == 'Books'"
        },
        {
            "data": [[0.4, 0.5, 0.6, ...]],
            "annsField": "sparse_vector",
            "limit": 50
        }
    ],
    rerank_strategy="weighted",
    rerank_params={"weights": [0.7, 0.3]},  # 70% dense, 30% sparse
    limit=10,
    output_fields=["title", "author", "price"],
    consistency_level="Bounded"
)
```

--------------------------------

### Suspend a Dedicated Cluster

Source: https://context7.com/zilliztech/zilliz-mcp-server/llms.txt

Use this function to suspend a dedicated cluster to reduce costs. The cluster will only incur storage costs while suspended.

```python
# MCP Tool Call
result = await suspend_cluster(cluster_id="inxx-xxxxxxxxxxxxxxx")
```

--------------------------------

### query_cluster_metrics

Source: https://context7.com/zilliztech/zilliz-mcp-server/llms.txt

Query performance metrics for a specific cluster including CPU computation, storage usage, request counts, and latency statistics.

```APIDOC
## query_cluster_metrics

### Description
Query performance metrics for a specific cluster including CPU computation, storage usage, request counts, and latency statistics.

### Parameters
#### Request Body
- **cluster_id** (string) - Required - The ID of the cluster.
- **period** (string) - Optional - ISO 8601 duration (e.g., PT1H).
- **start** (string) - Optional - Start timestamp.
- **end** (string) - Optional - End timestamp.
- **granularity** (string) - Required - Time interval (e.g., PT30S).
- **metric_queries** (array) - Required - List of metrics to query.

### Response
#### Success Response (200)
- **code** (integer) - Status code.
- **data** (object) - Metric results.
```

--------------------------------

### Insert Entities

Source: https://context7.com/zilliztech/zilliz-mcp-server/llms.txt

Inserts data entities into a specified collection. This operation supports inserting a single entity or multiple entities in a batch.

```APIDOC
## POST /api/collections/insert

### Description
Insert data entities into a collection. Supports single entity or batch insertion.

### Method
POST

### Endpoint
/api/collections/insert

### Parameters
#### Request Body
- **cluster_id** (string) - Required - The ID of the Zilliz cluster.
- **region_id** (string) - Required - The region where the cluster is located.
- **endpoint** (string) - Required - The API endpoint for the Zilliz cluster.
- **collection_name** (string) - Required - The name of the collection to insert data into.
- **data** (object|array) - Required - The data to insert. Can be a single entity object or an array of entity objects.
- **db_name** (string) - Optional - The name of the database (if applicable).

### Request Example (Batch Insertion)
```json
{
  "cluster_id": "inxx-xxxxxxxxxxxxxxx",
  "region_id": "gcp-us-west1",
  "endpoint": "https://inxx-xxxxxxxxxxxxxxx.api.gcp-us-west1.zillizcloud.com",
  "collection_name": "product_embeddings",
  "data": [
    {
      "vector": [0.1, 0.2, 0.3, ...],  
      "product_name": "Wireless Headphones",
      "category": "Electronics",
      "price": 99.99
    },
    {
      "vector": [0.4, 0.5, 0.6, ...],
      "product_name": "Laptop Stand",
      "category": "Accessories",
      "price": 49.99
    }
  ],
  "db_name": ""
}
```

### Request Example (Single Entity Insertion)
```json
{
  "cluster_id": "inxx-xxxxxxxxxxxxxxx",
  "region_id": "gcp-us-west1",
  "endpoint": "https://inxx-xxxxxxxxxxxxxxx.api.gcp-us-west1.zillizcloud.com",
  "collection_name": "product_embeddings",
  "data": {
    "vector": [0.7, 0.8, 0.9, ...],
    "product_name": "USB Cable",
    "category": "Accessories",
    "price": 9.99
  }
}
```

### Response
#### Success Response (200)
- **code** (integer) - The status code of the operation.
- **data** (object) - An object containing information about the insertion.
  - **insertCount** (integer) - The number of entities successfully inserted.
  - **insertIds** (array) - An array of IDs for the inserted entities.

#### Response Example
```json
{
  "code": 0,
  "data": {
    "insertCount": 2,
    "insertIds": [448300048035776800, 448300048035776801]
  }
}
```
```

--------------------------------

### Insert Entities into Collection

Source: https://context7.com/zilliztech/zilliz-mcp-server/llms.txt

Insert data entities into a collection. Supports both single entity and batch insertion. Automatic ID generation is used if `auto_id` is true during collection creation.

```python
# MCP Tool Call - Insert multiple entities
result = await insert_entities(
    cluster_id="inxx-xxxxxxxxxxxxxxx",
    region_id="gcp-us-west1",
    endpoint="https://inxx-xxxxxxxxxxxxxxx.api.gcp-us-west1.zillizcloud.com",
    collection_name="product_embeddings",
    data=[
        {
            "vector": [0.1, 0.2, 0.3, ...],  # 768-dimensional vector
            "product_name": "Wireless Headphones",
            "category": "Electronics",
            "price": 99.99
        },
        {
            "vector": [0.4, 0.5, 0.6, ...],
            "product_name": "Laptop Stand",
            "category": "Accessories",
            "price": 49.99
        }
    ],
    db_name=""  # Optional for free/serverless clusters
)
```

```python
# MCP Tool Call - Insert single entity
result = await insert_entities(
    cluster_id="inxx-xxxxxxxxxxxxxxx",
    region_id="gcp-us-west1",
    endpoint="https://inxx-xxxxxxxxxxxxxxx.api.gcp-us-west1.zillizcloud.com",
    collection_name="product_embeddings",
    data={
        "vector": [0.7, 0.8, 0.9, ...],
        "product_name": "USB Cable",
        "category": "Accessories",
        "price": 9.99
    }
)
```

--------------------------------

### Vector Search with Scalar Filter

Source: https://context7.com/zilliztech/zilliz-mcp-server/llms.txt

Enhance vector search by applying scalar filters to narrow down results based on metadata. Supports conditions on fields like category and price.

```python
result = await search(
    cluster_id="inxx-xxxxxxxxxxxxxxx",
    region_id="gcp-us-west1",
    endpoint="https://inxx-xxxxxxxxxxxxxxx.api.gcp-us-west1.zillizcloud.com",
    collection_name="product_embeddings",
    data=[[0.1, 0.2, 0.3, ...]],
    anns_field="vector",
    limit=5,
    filter="category == 'Electronics' and price < 100",
    output_fields=["product_name", "category", "price"],
    metric_type="COSINE"
)
```

--------------------------------

### Query Entities by Scalar Filter

Source: https://context7.com/zilliztech/zilliz-mcp-server/llms.txt

Retrieve entities based on scalar filter expressions without vector similarity. Useful for metadata-driven lookups.

```python
result = await query(
    cluster_id="inxx-xxxxxxxxxxxxxxx",
    region_id="gcp-us-west1",
    endpoint="https://inxx-xxxxxxxxxxxxxxx.api.gcp-us-west1.zillizcloud.com",
    collection_name="product_embeddings",
    filter="category == 'Electronics' and price > 50",
    output_fields=["product_name", "category", "price"],
    limit=100
)
```

--------------------------------

### suspend_cluster

Source: https://context7.com/zilliztech/zilliz-mcp-server/llms.txt

Suspend a dedicated cluster to save costs. The cluster will not incur computing costs when suspended—only storage costs apply.

```APIDOC
## suspend_cluster

### Description
Suspend a dedicated cluster to save costs. The cluster will not incur computing costs when suspended—only storage costs apply.

### Parameters
#### Request Body
- **cluster_id** (string) - Required - The ID of the cluster to suspend.

### Request Example
{
  "cluster_id": "inxx-xxxxxxxxxxxxxxx"
}

### Response
#### Success Response (200)
- **cluster_id** (string) - The ID of the cluster.
- **prompt** (string) - Confirmation message.

#### Response Example
{
  "cluster_id": "inxx-xxxxxxxxxxxxxxx",
  "prompt": "Successfully Submitted. The cluster will not incur any computing costs when suspended. You will only be billed for the storage costs during this time."
}
```

--------------------------------

### Search

Source: https://context7.com/zilliztech/zilliz-mcp-server/llms.txt

Performs a vector similarity search on a collection. This operation allows for optional scalar filtering and returns the most similar entities based on vector distance.

```APIDOC
## POST /api/collections/search

### Description
Perform vector similarity search on a collection with optional scalar filtering. Returns the most similar entities based on vector distance.

### Method
POST

### Endpoint
/api/collections/search

### Parameters
#### Request Body
- **cluster_id** (string) - Required - The ID of the Zilliz cluster.
- **region_id** (string) - Required - The region where the cluster is located.
- **endpoint** (string) - Required - The API endpoint for the Zilliz cluster.
- **collection_name** (string) - Required - The name of the collection to search within.
- **query_vectors** (array) - Required - An array of query vectors for the search.
- **limit** (integer) - Required - The maximum number of results to return.
- **filter** (string) - Optional - A filter expression to apply to the search.
- **output_fields** (array) - Optional - A list of fields to include in the search results.
- **params** (object) - Optional - Search parameters, such as "metric_type" and "params" for index.

### Request Example
```json
{
  "cluster_id": "inxx-xxxxxxxxxxxxxxx",
  "region_id": "gcp-us-west1",
  "endpoint": "https://inxx-xxxxxxxxxxxxxxx.api.gcp-us-west1.zillizcloud.com",
  "collection_name": "product_embeddings",
  "query_vectors": [
    [0.1, 0.2, 0.3, ...] 
  ],
  "limit": 10,
  "filter": "price > 50",
  "output_fields": ["product_name", "price"],
  "params": {
    "metric_type": "COSINE",
    "params": {"radius": 0.5}
  }
}
```

### Response
#### Success Response (200)
- **code** (integer) - The status code of the operation.
- **data** (object) - An object containing the search results.
  - **results** (array) - An array of search result objects, each containing matched entities and their scores.

#### Response Example
```json
{
  "code": 0,
  "data": {
    "results": [
      {
        "entities": [
          {
            "product_name": "Wireless Headphones",
            "price": 99.99,
            "vector": [0.1, 0.2, 0.3, ...]
          }
        ],
        "score": 0.95
      }
    ]
  }
}
```
```

--------------------------------

### Delete Entities from Collection

Source: https://context7.com/zilliztech/zilliz-mcp-server/llms.txt

Delete entities from a collection using filter expressions or primary key conditions. You can also specify a partition name to delete from a specific partition.

```python
# MCP Tool Call - Delete by filter expression
result = await delete_entities(
    cluster_id="inxx-xxxxxxxxxxxxxxx",
    region_id="gcp-us-west1",
    endpoint="https://inxx-xxxxxxxxxxxxxxx.api.gcp-us-west1.zillizcloud.com",
    collection_name="product_embeddings",
    filter="category == 'Discontinued'",
    db_name=""
)
```

```python
# MCP Tool Call - Delete by primary key
result = await delete_entities(
    cluster_id="inxx-xxxxxxxxxxxxxxx",
    region_id="gcp-us-west1",
    endpoint="https://inxx-xxxxxxxxxxxxxxx.api.gcp-us-west1.zillizcloud.com",
    collection_name="product_embeddings",
    filter="id in [448300048035776800, 448300048035776801]"
)
```

```python
# MCP Tool Call - Delete from specific partition
result = await delete_entities(
    cluster_id="inxx-xxxxxxxxxxxxxxx",
    region_id="gcp-us-west1",
    endpoint="https://inxx-xxxxxxxxxxxxxxx.api.gcp-us-west1.zillizcloud.com",
    collection_name="product_embeddings",
    filter="price < 10",
    partition_name="low_value_products"
)
```