An idiomatic C++ client library for the Cloud Vision API. Integrate Google Vision features, including image labeling, face, logo, and landmark detection, optical character recognition (OCR), and detection of explicit content, into applications.
While this library is GA, please note Google Cloud C++ client libraries do not follow Semantic Versioning.
Quickstart
The following shows the code that you'll run in the google/cloud/vision/quickstart/ directory, which should give you a taste of the Cloud Vision API C++ client library API.
#include "google/cloud/vision/v1/image_annotator_client.h"
#include <iostream>
int main(int argc, char* argv[]) try {
auto constexpr kDefaultUri =
"gs://cloud-samples-data/vision/label/wakeupcat.jpg";
if (argc > 2) {
std::cerr << "Usage: " << argv[0] << " [gcs-uri]\n"
<< " The gcs-uri must be in gs://... format. It defaults to "
<< kDefaultUri << "\n";
return 1;
}
auto uri = std::string{argc == 2 ? argv[1] : kDefaultUri};
namespace vision = ::google::cloud::vision_v1;
auto client =
vision::ImageAnnotatorClient(vision::MakeImageAnnotatorConnection());
// Define the image we want to annotate
google::cloud::vision::v1::Image image;
image.mutable_source()->set_image_uri(uri);
// Create a request to annotate this image with Request text annotations for a
// file stored in GCS.
google::cloud::vision::v1::AnnotateImageRequest request;
*request.mutable_image() = std::move(image);
request.add_features()->set_type(
google::cloud::vision::v1::Feature::TEXT_DETECTION);
google::cloud::vision::v1::BatchAnnotateImagesRequest batch_request;
*batch_request.add_requests() = std::move(request);
auto batch = client.BatchAnnotateImages(batch_request);
if (!batch) throw std::move(batch).status();
// Find the longest annotation and print it
auto result = std::string{};
for (auto const& response : batch->responses()) {
for (auto const& annotation : response.text_annotations()) {
if (result.size() < annotation.description().size()) {
result = annotation.description();
}
}
}
std::cout << "The image contains this text: " << result << "\n";
return 0;
} catch (google::cloud::Status const& status) {
std::cerr << "google::cloud::Status thrown: " << status << "\n";
return 1;
}
Main classes
This library offers multiple *Client classes, which are listed below. Each one of these classes exposes all the RPCs for a gRPC service as member functions of the class. This library groups multiple gRPC services because they are part of the same product or are often used together. A typical example may be the administrative and data plane operations for a single product.
The library also has other classes that provide helpers, configuration parameters, and infrastructure to mock the *Client classes when testing your application.
The library automatically retries requests that fail with transient errors, and uses exponential backoff to backoff between retries. Application developers can override the default policies.
More Information
Error Handling - describes how the library reports errors.
[[["Easy to understand","easyToUnderstand","thumb-up"],["Solved my problem","solvedMyProblem","thumb-up"],["Other","otherUp","thumb-up"]],[["Missing the information I need","missingTheInformationINeed","thumb-down"],["Too complicated / too many steps","tooComplicatedTooManySteps","thumb-down"],["Out of date","outOfDate","thumb-down"],["Samples / code issue","samplesCodeIssue","thumb-down"],["Other","otherDown","thumb-down"]],["Last updated 2025-08-14 UTC."],[[["\u003cp\u003eThis webpage provides documentation for the Cloud Vision API C++ Client Library, which allows developers to integrate Google Vision features like image labeling, OCR, and content detection into their applications.\u003c/p\u003e\n"],["\u003cp\u003eThe library is generally available (GA), but it is important to note that Google Cloud C++ client libraries do not adhere to Semantic Versioning standards.\u003c/p\u003e\n"],["\u003cp\u003eThe latest release candidate version available is 2.37.0-rc, and there is a full list of versions ranging from 2.11.0 to 2.36.0 and a complete list of all library versions available.\u003c/p\u003e\n"],["\u003cp\u003eThe library includes \u003ccode\u003e*Client\u003c/code\u003e classes, such as \u003ccode\u003eImageAnnotatorClient\u003c/code\u003e and \u003ccode\u003eProductSearchClient\u003c/code\u003e, each offering all the RPCs for a gRPC service as member functions, and it automatically handles retries and exponential backoff for transient errors.\u003c/p\u003e\n"],["\u003cp\u003eDevelopers have the ability to customize various settings, such as error handling, default endpoints, and authentication credentials, to modify the library's behavior.\u003c/p\u003e\n"]]],[],null,["Version 2.12.0keyboard_arrow_down\n\n- [2.42.0-rc (latest)](/cpp/docs/reference/vision/latest)\n- [2.41.0](/cpp/docs/reference/vision/2.41.0)\n- [2.40.0](/cpp/docs/reference/vision/2.40.0)\n- [2.39.0](/cpp/docs/reference/vision/2.39.0)\n- [2.38.0](/cpp/docs/reference/vision/2.38.0)\n- [2.37.0](/cpp/docs/reference/vision/2.37.0)\n- [2.36.0](/cpp/docs/reference/vision/2.36.0)\n- [2.35.0](/cpp/docs/reference/vision/2.35.0)\n- [2.34.0](/cpp/docs/reference/vision/2.34.0)\n- [2.33.0](/cpp/docs/reference/vision/2.33.0)\n- [2.32.0](/cpp/docs/reference/vision/2.32.0)\n- [2.31.0](/cpp/docs/reference/vision/2.31.0)\n- [2.30.0](/cpp/docs/reference/vision/2.30.0)\n- [2.29.0](/cpp/docs/reference/vision/2.29.0)\n- [2.28.0](/cpp/docs/reference/vision/2.28.0)\n- [2.27.0](/cpp/docs/reference/vision/2.27.0)\n- [2.26.0](/cpp/docs/reference/vision/2.26.0)\n- [2.25.1](/cpp/docs/reference/vision/2.25.1)\n- [2.24.0](/cpp/docs/reference/vision/2.24.0)\n- [2.23.0](/cpp/docs/reference/vision/2.23.0)\n- [2.22.1](/cpp/docs/reference/vision/2.22.1)\n- [2.21.0](/cpp/docs/reference/vision/2.21.0)\n- [2.20.0](/cpp/docs/reference/vision/2.20.0)\n- [2.19.0](/cpp/docs/reference/vision/2.19.0)\n- [2.18.0](/cpp/docs/reference/vision/2.18.0)\n- [2.17.0](/cpp/docs/reference/vision/2.17.0)\n- [2.16.0](/cpp/docs/reference/vision/2.16.0)\n- [2.15.1](/cpp/docs/reference/vision/2.15.1)\n- [2.14.0](/cpp/docs/reference/vision/2.14.0)\n- [2.13.0](/cpp/docs/reference/vision/2.13.0)\n- [2.12.0](/cpp/docs/reference/vision/2.12.0)\n- [2.11.0](/cpp/docs/reference/vision/2.11.0) \n\nCloud Vision API C++ Client Library\n===================================\n\nAn idiomatic C++ client library for the [Cloud Vision API](https://cloud.google.com/vision). Integrate Google Vision features, including image labeling, face, logo, and landmark detection, optical character recognition (OCR), and detection of explicit content, into applications.\n\nWhile this library is **GA** , please note Google Cloud C++ client libraries do **not** follow [Semantic Versioning](https://semver.org/).\n\n### Quickstart\n\nThe following shows the code that you'll run in the `google/cloud/vision/quickstart/` directory, which should give you a taste of the Cloud Vision API C++ client library API. \n\n #include \"google/cloud/vision/v1/image_annotator_client.h\"\n #include \u003ciostream\u003e\n\n int main(int argc, char* argv[]) try {\n auto constexpr kDefaultUri =\n \"gs://cloud-samples-data/vision/label/wakeupcat.jpg\";\n if (argc \u003e 2) {\n std::cerr \u003c\u003c \"Usage: \" \u003c\u003c argv[0] \u003c\u003c \" [gcs-uri]\\n\"\n \u003c\u003c \" The gcs-uri must be in gs://... format. It defaults to \"\n \u003c\u003c kDefaultUri \u003c\u003c \"\\n\";\n return 1;\n }\n auto uri = std::string{argc == 2 ? argv[1] : kDefaultUri};\n\n namespace vision = ::google::cloud::vision_v1;\n auto client =\n vision::ImageAnnotatorClient(vision::MakeImageAnnotatorConnection());\n\n // Define the image we want to annotate\n google::cloud::vision::v1::Image image;\n image.mutable_source()-\u003eset_image_uri(uri);\n // Create a request to annotate this image with Request text annotations for a\n // file stored in GCS.\n google::cloud::vision::v1::AnnotateImageRequest request;\n *request.mutable_image() = std::move(image);\n request.add_features()-\u003eset_type(\n google::cloud::vision::v1::Feature::TEXT_DETECTION);\n\n google::cloud::vision::v1::BatchAnnotateImagesRequest batch_request;\n *batch_request.add_requests() = std::move(request);\n auto batch = client.BatchAnnotateImages(batch_request);\n if (!batch) throw std::move(batch).status();\n\n // Find the longest annotation and print it\n auto result = std::string{};\n for (auto const& response : batch-\u003eresponses()) {\n for (auto const& annotation : response.text_annotations()) {\n if (result.size() \u003c annotation.description().size()) {\n result = annotation.description();\n }\n }\n }\n std::cout \u003c\u003c \"The image contains this text: \" \u003c\u003c result \u003c\u003c \"\\n\";\n\n return 0;\n } catch (google::cloud::Status const& status) {\n std::cerr \u003c\u003c \"google::cloud::Status thrown: \" \u003c\u003c status \u003c\u003c \"\\n\";\n return 1;\n }\n\n### Main classes\n\nThis library offers multiple `*Client` classes, which are listed below. Each one of these classes exposes all the RPCs for a gRPC `service` as member functions of the class. This library groups multiple gRPC services because they are part of the same product or are often used together. A typical example may be the administrative and data plane operations for a single product.\n\nThe library also has other classes that provide helpers, configuration parameters, and infrastructure to mock the `*Client` classes when testing your application.\n\n- [`vision_v1::ImageAnnotatorClient`](/cpp/docs/reference/vision/2.12.0/classgoogle_1_1cloud_1_1vision__v1_1_1ImageAnnotatorClient)\n- [`vision_v1::ProductSearchClient`](/cpp/docs/reference/vision/2.12.0/classgoogle_1_1cloud_1_1vision__v1_1_1ProductSearchClient)\n\n### Retry, Backoff, and Idempotency Policies.\n\nThe library automatically retries requests that fail with transient errors, and uses [exponential backoff](https://en.wikipedia.org/wiki/Exponential_backoff) to backoff between retries. Application developers can override the default policies.\n\n### More Information\n\n- [Error Handling](https://cloud.google.com/cpp/docs/reference/common/latest/common-error-handling.html) - describes how the library reports errors.\n- [How to Override the Default Endpoint](/cpp/docs/reference/vision/2.12.0/vision-override-endpoint) - describes how to override the default endpoint.\n- [How to Override the Authentication Credentials](/cpp/docs/reference/vision/2.12.0/vision-override-authentication) - describes how to change the authentication credentials used by the library.\n- [Environment Variables](/cpp/docs/reference/vision/2.12.0/vision-env) - describes environment variables that can configure the behavior of the library."]]