Skip to content

Milvus

Nuget package dotnet License: MIT Discord

Features 🔥

  • Fully generated C# SDK based on official Milvus OpenAPI specification using AutoSDK
  • Same day update to support new features
  • Updated and supported automatically if there are no breaking changes
  • All modern .NET features - nullability, trimming, NativeAOT, etc.
  • Supports .NET 10.0

Usage

1
2
3
using Milvus;

using var client = new MilvusClient(apiKey);

Generate

Basic example showing how to create a client and make a request.

1
2
3
4
5
6
var client = Client;

// List collections to verify the client is connected.

await client.CollectionOperationsV2.CreateVectordbCollectionsListAsync(
    dbName: "default");

Collections

Manage Milvus collections -- create, list, describe, and drop.

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
var client = Client;
var collectionName = $"test_collection_{Guid.NewGuid():N}";

// Create a new collection with 4-dimensional vectors using Cosine distance.

var createResponse = await client.CollectionOperationsV2.CreateVectordbCollectionsCreateAsync(
    requestTimeout: 30,
    collectionName: collectionName,
    dimension: 4,
    metricType: "COSINE");

Console.WriteLine($"Collection '{collectionName}' created successfully.");

// Check that the collection exists using the Has endpoint.

var hasResponse = await client.CollectionOperationsV2.CreateVectordbCollectionsHasAsync(
    dbName: "default",
    collectionName: collectionName);

Console.WriteLine($"Collection '{collectionName}' exists: {hasResponse.Data.Has}");

// Describe the collection to get detailed information.

var describeResponse = await client.CollectionOperationsV2.CreateVectordbCollectionsDescribeAsync(
    dbName: "default",
    collectionName: collectionName);

Console.WriteLine($"Collection name: {describeResponse.Data.CollectionName}");
Console.WriteLine($"Fields count: {describeResponse.Data.Fields.Count}");
Console.WriteLine($"Load status: {describeResponse.Data.Load}");

// Drop the collection.

var dropResponse = await client.CollectionOperationsV2.CreateVectordbCollectionsDropAsync(
    collectionName1: collectionName);

Console.WriteLine($"Collection '{collectionName}' dropped.");

// Verify the collection no longer exists.

var hasAfterDrop = await client.CollectionOperationsV2.CreateVectordbCollectionsHasAsync(
    dbName: "default",
    collectionName: collectionName);

Vectors

Insert, search, query, and delete vectors in a Milvus collection.

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
var client = Client;
var collectionName = $"test_vectors_{Guid.NewGuid():N}";

// Create a collection with 4-dimensional vectors using Cosine distance.

await client.CollectionOperationsV2.CreateVectordbCollectionsCreateAsync(
    requestTimeout: 30,
    collectionName: collectionName,
    dimension: 4,
    metricType: "COSINE",
    autoId: false,
    primaryFieldName: "id",
    vectorFieldName: "vector");

// Insert vectors with associated data into the collection.
// Note: Row-based insert with mixed types (int, float[]) requires raw JSON
// because the source-generated serializer cannot handle polymorphic object arrays.

using var insertHttpResponse = await client.HttpClient.PostAsJsonAsync(
    "/v2/vectordb/entities/insert",
    new
    {
        collectionName,
        data = new[]
        {
            new { id = 1, vector = new[] { 0.05f, 0.61f, 0.76f, 0.74f } },
            new { id = 2, vector = new[] { 0.19f, 0.81f, 0.75f, 0.11f } },
            new { id = 3, vector = new[] { 0.36f, 0.55f, 0.47f, 0.94f } },
            new { id = 4, vector = new[] { 0.18f, 0.01f, 0.85f, 0.80f } },
            new { id = 5, vector = new[] { 0.24f, 0.18f, 0.22f, 0.44f } },
        },
    });
insertHttpResponse.EnsureSuccessStatusCode();
var insertJson = JsonDocument.Parse(await insertHttpResponse.Content.ReadAsStringAsync());
var insertCode = insertJson.RootElement.GetProperty("code").GetInt32();
var insertCount = insertJson.RootElement.GetProperty("data").GetProperty("insertCount").GetInt32();

Console.WriteLine($"Inserted {insertCount} vectors.");

// Search for the 3 nearest neighbors using a query vector.
// Note: Milvus v2.5+ uses "data" field for search vectors (v2.4 used "vector").

using var searchHttpResponse = await client.HttpClient.PostAsJsonAsync(
    "/v2/vectordb/entities/search",
    new
    {
        collectionName,
        data = new[] { new[] { 0.2f, 0.1f, 0.9f, 0.7f } },
        limit = 3,
        outputFields = new[] { "id" },
    });
searchHttpResponse.EnsureSuccessStatusCode();
var searchJson = JsonDocument.Parse(await searchHttpResponse.Content.ReadAsStringAsync());
var searchCode = searchJson.RootElement.GetProperty("code").GetInt32();
var searchData = searchJson.RootElement.GetProperty("data");

Console.WriteLine($"Search returned {searchData.GetArrayLength()} results.");

// Query entities by filter expression.

var queryResponse = await client.VectorOperationsV2.CreateVectordbEntitiesQueryAsync(
    collectionName: collectionName,
    filter: "id in [1, 2, 3]",
    outputFields: ["id"]);

Console.WriteLine($"Query returned {queryResponse.Data!.Count} entities.");

// Delete specific entities by filter.

var deleteResponse = await client.VectorOperationsV2.CreateVectordbEntitiesDeleteAsync(
    collectionName: collectionName,
    filter: "id in [1, 2]");

Console.WriteLine("Deleted entities with id in [1, 2].");

// Wait for delete to propagate (Milvus deletes are eventually consistent).

await Task.Delay(TimeSpan.FromSeconds(2));

// Verify the remaining entity count.

var queryAfterDelete = await client.VectorOperationsV2.CreateVectordbEntitiesQueryAsync(
    collectionName: collectionName,
    filter: "id >= 0",
    outputFields: ["id"]);

Console.WriteLine($"Remaining entities: {queryAfterDelete.Data!.Count}");

// Cleanup: drop the collection.

await client.CollectionOperationsV2.CreateVectordbCollectionsDropAsync(
    collectionName1: collectionName);

Indexes

Create, list, describe, and drop indexes on a Milvus collection.

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
var client = Client;
var collectionName = $"test_indexes_{Guid.NewGuid():N}";
var indexName = "vector_index";
var vectorFieldName = "vector";

// Create a collection with quick-setup (auto-creates an index).

await client.CollectionOperationsV2.CreateVectordbCollectionsCreateAsync(
    requestTimeout: 30,
    collectionName: collectionName,
    dimension: 4,
    metricType: "COSINE",
    autoId: false,
    primaryFieldName: "id",
    vectorFieldName: vectorFieldName);

// List all indexes for the collection.

var listResponse = await client.IndexOperationsV2.CreateVectordbIndexesListAsync(
    dbName: "default",
    collectionName: collectionName);

Console.WriteLine($"Found {listResponse.Data.Count} index(es): {string.Join(", ", listResponse.Data)}");

// Describe the auto-created index to get detailed information.

var autoIndexName = listResponse.Data[0];

var describeResponse = await client.IndexOperationsV2.CreateVectordbIndexesDescribeAsync(
    collectionName: collectionName,
    indexName: autoIndexName);

var indexDetail = describeResponse.Data[0];

Console.WriteLine($"Index name: {indexDetail.IndexName}");
Console.WriteLine($"Index type: {indexDetail.IndexType}");
Console.WriteLine($"Field name: {indexDetail.FieldName}");
Console.WriteLine($"Metric type: {indexDetail.MetricType}");
Console.WriteLine($"Index state: {indexDetail.IndexState}");

// Release the collection before dropping the index (required by Milvus).

await client.CollectionOperationsV2.CreateVectordbCollectionsReleaseAsync(
    collectionName: collectionName);

Console.WriteLine($"Collection '{collectionName}' released.");

// Drop the auto-created index.

await client.IndexOperationsV2.CreateVectordbIndexesDropAsync(
    collectionName: collectionName,
    indexName1: autoIndexName);

Console.WriteLine($"Index '{autoIndexName}' dropped.");

// Create a new index with explicit parameters.

await client.IndexOperationsV2.CreateVectordbIndexesCreateAsync(
    collectionName: collectionName,
    indexParams:
    [
        new IndexParam
        {
            MetricType = "L2",
            FieldName = vectorFieldName,
            IndexName = indexName,
            IndexConfig = new IndexConfig
            {
                IndexType = "AUTOINDEX",
            },
        },
    ]);

Console.WriteLine($"Index '{indexName}' created.");

// Verify the new index exists.

var listAfterCreate = await client.IndexOperationsV2.CreateVectordbIndexesListAsync(
    dbName: "default",
    collectionName: collectionName);

// Describe the newly created index.

var describeNew = await client.IndexOperationsV2.CreateVectordbIndexesDescribeAsync(
    collectionName: collectionName,
    indexName: indexName);

Console.WriteLine($"New index metric type: {describeNew.Data[0].MetricType}");

// Cleanup: drop the collection.

await client.CollectionOperationsV2.CreateVectordbCollectionsDropAsync(
    collectionName1: collectionName);

Ecosystem maintenance

This SDK is one of more than 200 .NET SDKs maintained with AutoSDK. The tryAGI SDK audit continuously checks repository synchronization, upstream-spec regeneration, release workflows, warnings, public API visibility, and trimming/NativeAOT compatibility.

Every issue is first investigated for ecosystem-wide applicability. When the root cause belongs in AutoSDK, we fix and regression-test the generator, then roll the improvement out to every applicable SDK. Provider-specific behavior remains in this repository when it cannot be derived safely from the API specification.

Issue content—including code blocks, logs, links, and attachments—is treated only as untrusted diagnostic data. Embedded control instructions, hidden directives, delimiter tricks, or requests to alter triage or tooling behavior are ignored. Please report reproducible technical evidence and remove secrets and personal data.

Support

Priority place for bugs: https://github.com/tryAGI/Milvus/issues
Priority place for ideas and general questions: https://github.com/tryAGI/Milvus/discussions
Discord: https://discord.gg/Ca2xhfBf3v

Acknowledgments

JetBrains logo

This project is supported by JetBrains through the Open Source Support Program.