# Vertex AI Samples: what Google's sample repository actually gives you > The GoogleCloudPlatform/vertex-ai-samples repository holds notebooks, sample apps and agent skills for Vertex AI. It is a teaching resource tied to a paid cloud platform, not a library you install, and the disclaimer says so. **GoogleCloudPlatform/vertex-ai-samples** — Notebooks, code samples, sample apps, and other resources that demonstrate how to use, develop and manage machine learning and generative AI workflows using Google Cloud Vertex AI. - Repository: https://github.com/GoogleCloudPlatform/vertex-ai-samples - Website: https://cloud.google.com/vertex-ai - Stars: 791 · Forks: 315 - Language: Jupyter Notebook - License: Apache-2.0 - Published: 2026-09-10 · Updated: 2026-09-10 · Language: en - Canonical page: https://hysenlabs.com/projects/googlecloudplatform-vertex-ai-samples ## What vertex-ai-samples is for, and who it is not for This is a sample repository, not a library. It exists to show how Vertex AI services are used: the README describes it as notebooks, code samples, sample apps, skills and other resources that demonstrate how to use, develop and manage machine learning and generative AI workflows on Vertex AI. Nothing here is published as a package you depend on, and there is no release history in the repository metadata. You consume it by reading and running the files. The audience is narrow and specific. You need a Google Cloud project before anything works; the README's Get started section says so directly and links to the free trial and to the setup guide for a project and development environment. If you are evaluating whether to move ML workloads onto Vertex AI, the notebooks are a reasonable way to see the service surface before committing. If you want a self-hosted or cloud-neutral stack, this repository has nothing for you, because every sample assumes Vertex AI endpoints and Google Cloud services. The repository also carries an explicit disclaimer: it is not an officially supported Google product and the code is for demonstrative purposes only. Treat that as the operating assumption for everything below. ## How the repository is organised: official notebooks, community notebooks, skills The README prints the directory tree, and it is the clearest statement of the architecture. Under notebooks there are two branches: official, which the README describes as notebooks demonstrating use of each Vertex AI service, and community, for notebooks contributed by the community. Official is further split by service, with automl, custom and others listed, while community includes model_garden and similar folders. Outside notebooks sit community-content for contributed sample code and tutorials, docs for deep-dive documentation and advanced setup guides, and skills. The skills directory is the part that does not look like the rest. The tree shows a vertex-ai/SKILL.md described as the entry point that routes across capabilities, plus genai-sdk for Gemini API usage with the Gen AI SDK in Python, JS/TS, Go, Java and C#, vertex-deploy for deploying models to Endpoints, vertex-inference for inferencing with GenAI models, and vertex-tuning as a secondary router with gemini/ and open-model/ subfolders. That is a different artefact type: structured instructions an agent can follow, not a notebook a human runs cell by cell. If your interest is agent tooling rather than tutorials, that directory is the reason to look here. Note what the folder names imply about scope. Official notebooks are organised one service per folder, so a sample for Feature Store does not teach you Model Registry. You will be jumping between directories rather than following a single learning path. ## Running a first sample: project setup, Colab, and the notebook header There is no install command for this repository as a whole. The README's Get started section gives the prerequisite instead: you must have a Google Cloud project, and if you do not have one it points to the Google Cloud free trial. Once a project exists, it links to the Vertex AI documentation page for setting up a project and a development environment. That page, not this repository, is where the real setup steps live. The workflow the README describes is to explore the repository and follow the links in the header section of each notebook. Those header links are the mechanism: each notebook offers to open and run in Colab, in Colab Enterprise, in Vertex AI Workbench, or to be viewed on GitHub. So the practical first step is to pick a notebook and use its own header rather than cloning anything. If you prefer to work from a clone, the repository is a plain Git checkout and the notebooks are the files: ```bash git clone https://github.com/GoogleCloudPlatform/vertex-ai-samples.git cd vertex-ai-samples ls notebooks/official ``` You should see the service folders listed in the README tree, such as automl, custom and pipelines. From there, open a notebook from notebooks/official in your chosen environment. The README does not document a CLI, an environment file or a dependency manifest for the repository, so any package installation happens inside the individual notebook, not at the repository root. ## Where the samples stop being useful The disclaimer is the first limitation and the most important one. The README states plainly that this is not an officially supported Google product and that the code is for demonstrative purposes only. There is no support commitment attached to a notebook, and nothing in the README promises that a sample keeps working as Vertex AI changes. The second limitation is structural. The README has no changelog and the metadata shows no releases, so there is no version of this repository to pin. A notebook that worked against a service six months ago carries no compatibility guarantee, and the repository does not tell you which service version it was written for. That is a real cost if you plan to lift code out of a sample into something you maintain. The third is scope. The README redirects generative AI notebook samples elsewhere, pointing readers to the separate GoogleCloudPlatform/generative-ai repository for more Vertex AI Generative AI notebook samples. So if your work is Gemini and generative AI, this repository is not the primary source even though it contains genai-related skills and a model_garden folder. Finally, the cost model is inverted from most open source. The code is Apache-2.0 and free to take, but running any of it consumes a Google Cloud project and the Vertex AI services behind it. The samples are cheap; the platform they demonstrate is not. ## vertex-ai-samples compared with the generative-ai samples repository The honest alternative is not a competing vendor. It is the other Google repository the README itself recommends: GoogleCloudPlatform/generative-ai. The README sends readers there for more Vertex AI Generative AI notebook samples, which tells you the split. This repository covers the broader Vertex AI surface, with official notebooks grouped by service such as automl, custom, feature_store, datasets, prediction, model_registry, explainable_ai, ml_metadata and pipelines, plus the skills directory. The generative-ai repository is where generative AI notebook work is concentrated. The difference in approach is one of breadth against depth in a single area. If you are working on classical ML on Vertex AI, model deployment, pipelines, feature serving or explainability, this repository is the one whose folder structure matches your task. If your work is prompt-driven generative AI, the README is effectively telling you that the other repository is the better starting point, and you should follow that rather than digging through community notebooks here. A second alternative worth naming is the official Vertex AI documentation itself. The README links to it under References alongside the Jupyter notebook tutorials page. Documentation describes the supported contract; these samples demonstrate usage. When the two disagree, the documentation is the one that governs what the service does. ## Licence, contribution and the cost of staying current The repository is Apache-2.0, which permits commercial use and modification under its terms. That covers the sample code. It does not cover the Vertex AI services the samples call, which are governed by Google Cloud's own terms and billed separately. The disclaimer that this is not an officially supported Google product also means the licence gives you the code, not a support relationship. This is a description of what the repository states, not legal advice; read the LICENSE file and Google Cloud's terms for your own situation. Contributions go through the Contributing Guide linked from the README, and feedback or bug reports go to the Issues page. The repository has a CODEOWNERS file and a SECURITY.md at the top level, and a renovate.json, which suggests some automated dependency update activity, though the README does not describe a release process or a deprecation policy for notebooks. The upgrade cost is therefore manual and per notebook. Because there is no versioned artefact, keeping a sample working means re-running it against current Vertex AI and fixing what broke. The repository's last push was on 2026-09-04, so it is receiving changes, but that tells you nothing about whether any specific notebook you copied is still accurate. Budget for re-verification each time you revisit a sample. ## Conclusion Adopt vertex-ai-samples if you already have a Google Cloud project and want runnable starting points for Vertex AI services such as Pipelines, Feature Store, Model Registry or AutoML, or if you want the skills/ directory as agent instructions. Do not adopt it if you need a supported product, a versioned package, or anything that runs outside Google Cloud; the README states it is not an officially supported Google product and is for demonstrative purposes only. Before building on a notebook, open it and check the header links and which services it calls, because that is where the samples point you at Colab, Colab Enterprise or Vertex AI Workbench, and confirm current Vertex AI behaviour against the official documentation rather than the notebook text. ## FAQ ### What is Vertex AI used for? According to the README, Vertex AI is a fully managed, unified AI development platform for building and using generative AI. The samples in this repository demonstrate workflows across that platform, including AutoML, custom training, prediction, pipelines, Feature Store, Model Registry and explainability. ### Is Vertex AI the same as Gemini? The repository does not treat them as the same thing. Vertex AI is described as the platform, and Gemini appears as one of the models available on it, listed in the Model Garden row alongside Gemma, Llama 3 and Claude 3. The README does not state that the two names are interchangeable. ### What happened to Vertex AI? Nothing in the repository indicates a change or discontinuation. The README describes Vertex AI in the present tense as a fully managed, unified AI development platform, and the repository's last push was on 2026-09-04. ## Sources - [GoogleCloudPlatform/vertex-ai-samples on GitHub](https://github.com/GoogleCloudPlatform/vertex-ai-samples) - [Issues](https://github.com/GoogleCloudPlatform/vertex-ai-samples/issues) - [License: Apache-2.0](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/LICENSE) - [Project website](https://cloud.google.com/vertex-ai) - [README](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/README.md) --- Hysen Labs editorial analysis, written from the project's own repository and release notes. Cite the canonical page: https://hysenlabs.com/projects/googlecloudplatform-vertex-ai-samples