mostly-ai · mplatzer · Apr 27, 2026 · Apr 25, 2026 · Apr 25, 2026 · Apr 25, 2026
diff --git a/.github/ISSUE_TEMPLATE/config.yml b/.github/ISSUE_TEMPLATE/config.yml
@@ -1,6 +1 @@
 blank_issues_enabled: false  # Disables the option to create blank issues
-
-contact_links:
-  - name: Contact Support
-    url: mailto:support@mostly.ai
-    about: For any support-related queries, please email us directly at support@mostly.ai.
diff --git a/.github/workflows/run-tests-cpu.yaml b/.github/workflows/run-tests-cpu.yaml
@@ -40,37 +40,3 @@ jobs:
           uv run --no-sync pytest -vv tests/_data/unit
           uv run --no-sync pytest -vv tests/_local/unit
           uv run --no-sync pytest -vv tests/test_domain.py
-
-  run-test-cpu-client:
-    if: false
-    runs-on: ubuntu-latest
-    permissions:
-      contents: read
-      packages: write
-    steps:
-      - name: Setup | Checkout
-        uses: actions/checkout@8e8c483db84b4bee98b60c0593521ed34d9990e8 # v6.0.1
-        with:
-          fetch-depth: 1
-          submodules: 'true'
-
-      - name: Setup | uv
-        uses: astral-sh/setup-uv@ed21f2f24f8dd64503750218de024bcf64c7250a # v7.1.5
-        with:
-          enable-cache: false
-          python-version: '3.11'
-
-      - name: Setup | Dependencies
-        run: |
-          uv sync --frozen --only-group dev
-          uv pip install --index-strategy unsafe-first-match torch==2.9.1+cpu torchvision==0.24.1+cpu ".[local]" --extra-index-url https://download.pytorch.org/whl/cpu
-
-      - name: Test | End-to-End (Client mode only)
-        env:
-          MOSTLY_API_KEY: ${{ secrets.E2E_CLIENT_MOSTLY_API_KEY }}
-          MOSTLY_BASE_URL: ${{ secrets.E2E_CLIENT_MOSTLY_BASE_URL }}
-          E2E_CLIENT_S3_ACCESS_KEY: ${{ secrets.E2E_CLIENT_S3_ACCESS_KEY }}
-          E2E_CLIENT_S3_SECRET_KEY: ${{ secrets.E2E_CLIENT_S3_SECRET_KEY }}
-          E2E_CLIENT_S3_BUCKET: ${{ secrets.E2E_CLIENT_S3_BUCKET }}
-        run: |
-          uv run --no-sync pytest -vv tests/_local/end_to_end -k 'client and mode'
diff --git a/Dockerfile b/Dockerfile
@@ -66,10 +66,9 @@ RUN uv pip install torch==2.9.1 torchvision==0.24.1 --torch-backend=cpu
 COPY mostlyai ./mostlyai
 COPY README.md ./
 RUN uv pip install -e .
-COPY ./tools/docker_entrypoint.py /app/entrypoint.py
 
 USER nonroot
 
 EXPOSE 8080
 ENTRYPOINT [ "uv", "run", "--no-sync", "--project", "/app", "--"]
-CMD ["/app/entrypoint.py"]
+CMD ["python", "-m", "mostlyai.sdk._local.docker_entrypoint"]
diff --git a/Makefile b/Makefile
@@ -1,42 +1,8 @@
-# Internal Variables
-PUBLIC_API_FULL_URL = https://raw.githubusercontent.com/mostly-ai/mostly-openapi/refs/heads/main/public-api.yaml
-PUBLIC_API_OUTPUT_PATH = mostlyai/sdk/domain.py
-
 # Targets
 .PHONY: help
 help: ## show definition of each function
 	@awk 'BEGIN {FS = ":.*?## "} /^[a-zA-Z_-]+:.*?## / {printf "\033[36m%-25s\033[0m %s\n", $$1, $$2}' $(MAKEFILE_LIST)
 
-.PHONY: gen-public-model
-gen-public-model: ## build pydantic models for public api
-	@echo "Updating custom Jinja2 templates"
-	python tools/extend_model.py
-	@echo "Generating Pydantic models from $(PUBLIC_API_FULL_URL)"
-	datamodel-codegen --url $(PUBLIC_API_FULL_URL) $(COMMON_OPTIONS)
-	#datamodel-codegen --input ../mostly-app-v2/public-api/public-api.yaml $(COMMON_OPTIONS)
-	python tools/postproc_domain.py
-	# run pre-commit hooks to add license and lint the generated code; ignore the exit code to avoid confusion
-	uv run --no-sync pre-commit run --all-files > /dev/null 2>&1 || true
-
-# Common options for both targets
-COMMON_OPTIONS = \
-	--input-file-type openapi \
-	--output $(PUBLIC_API_OUTPUT_PATH) \
-	--snake-case-field \
-	--target-python-version 3.11 \
-	--use-schema-description \
-	--use-union-operator \
-	--use-standard-collections \
-	--field-constraints \
-	--collapse-root-models \
-	--use-one-literal-as-default \
-	--enum-field-as-literal one \
-	--use-subclass-enum \
-	--set-default-enum-member \
-	--output-model-type pydantic_v2.BaseModel \
-	--base-class mostlyai.sdk.client.base.CustomBaseModel \
-	--custom-template-dir tools/custom_template
-
 .PHONY: clean
 clean: ## Remove .gitignore files
 	git clean -fdX

diff --git a/README.md b/README.md
@@ -7,13 +7,12 @@
 [![PyPI - Python Version](https://img.shields.io/pypi/pyversions/mostlyai)](https://pypi.org/project/mostlyai/)
 [![GitHub stars](https://img.shields.io/github/stars/mostly-ai/mostlyai?style=social)](https://github.com/mostly-ai/mostlyai/stargazers)
 
-[Documentation](https://mostly-ai.github.io/mostlyai/) | [Technical White Paper](https://arxiv.org/abs/2508.00718) | [Usage Examples](https://mostly-ai.github.io/mostlyai/usage/) | [Free Cloud Service](https://app.mostly.ai/)
+[Documentation](https://mostly-ai.github.io/mostlyai/) | [Technical White Paper](https://arxiv.org/abs/2508.00718) | [Usage Examples](https://mostly-ai.github.io/mostlyai/usage/)
 
 The **Synthetic Data SDK** is a Python toolkit for high-fidelity, privacy-safe **Synthetic Data**.
 
-- **LOCAL** mode trains and generates synthetic data locally on your own compute resources.
-- **CLIENT** mode connects to a remote MOSTLY AI platform for training & generating synthetic data there.
-- Generators, that were trained locally, can be easily imported to a platform for further sharing.
+- **LOCAL** mode trains and generates synthetic data locally on your own compute resources (default).
+- **CLIENT** mode connects to a remote SDK endpoint for training & generating synthetic data there.
 
 ## Overview
 
@@ -30,8 +29,6 @@ The SDK allows you to programmatically create, browse and manage 3 key resources
 | Live probe the generator on demand            | `df = mostly.probe(g, config)`    | [mostly.probe](https://mostly-ai.github.io/mostlyai/api_client/#mostlyai.sdk.client.api.MostlyAI.probe)       |
 | Connect to any data source within your org    | `c = mostly.connect(config)`      | [mostly.connect](https://mostly-ai.github.io/mostlyai/api_client/#mostlyai.sdk.client.api.MostlyAI.connect)   |
 
-https://github.com/user-attachments/assets/9e233213-a259-455c-b8ed-d1f1548b492f
-
 ## Key Features
 
 - **Broad Data Support**
@@ -62,10 +59,10 @@ https://github.com/user-attachments/assets/9e233213-a259-455c-b8ed-d1f1548b492f
 
 ## Quick Start <a href="https://colab.research.google.com/github/mostly-ai/mostlyai/blob/main/docs/tutorials/getting-started/getting-started.ipynb" target="_blank"><img src="https://img.shields.io/badge/Open%20in-Colab-blue?logo=google-colab" alt="Run on Colab"></a>
 
-Install the SDK via `pip` (see [Installation](#installation) for further details):
+Install the SDK for LOCAL mode (see [Installation](#installation) for further details):
 
 ```shell
-pip install -U mostlyai  # or 'mostlyai[local]' for LOCAL mode
+uv pip install -U 'mostlyai[local]'
 ```
 
 Generate synthetic samples using a pre-trained generator:
@@ -191,17 +188,9 @@ uv run --with jupyter jupyter lab
 
 </details>
 
-### CLIENT mode
-
-This is a light-weight installation for using the SDK in CLIENT mode only. It communicates to a MOSTLY AI platform to perform requested tasks. See e.g. [app.mostly.ai](https://app.mostly.ai/) for a free-to-use hosted version.
-
-```shell
-uv pip install -U mostlyai
-```
-
-### CLIENT + LOCAL mode
+### LOCAL mode (default)
 
-This is a full installation for using the SDK in both CLIENT and LOCAL mode. It includes all dependencies, incl. PyTorch, for training and generating synthetic data locally.
+This is the recommended installation for running the SDK in LOCAL mode (and optionally CLIENT mode as well). It includes all dependencies, incl. PyTorch, for training and generating synthetic data locally.
 
 ```shell
 uv pip install -U 'mostlyai[local]'
@@ -225,15 +214,23 @@ uv pip install --index-strategy unsafe-first-match -U torch==2.9.1+cpu torchvisi
 pip install -U torch==2.9.1+cpu torchvision==0.24.1+cpu 'mostlyai[local]' --extra-index-url https://download.pytorch.org/whl/cpu
 ```
 
+### CLIENT mode
+
+This is a light-weight installation for explicit CLIENT-only usage against a remote SDK endpoint.
+
+```shell
+uv pip install -U mostlyai
+```
+
 
 > **Note for Google Colab users**: Installing any of the local extras (`mostlyai[local]`, or `mostlyai[local-gpu]`) might need restarting the runtime after installation for the changes to take effect.
 
 ### Data Connectors
 
-Add any of the following extras for further data connectors support in LOCAL mode: `databricks`, `googlebigquery`, `hive`, `mssql`, `mysql`, `oracle`, `postgres`, `redshift`, `snowflake`. E.g.
+Add any of the following extras for further data connectors support in LOCAL mode: `googlebigquery`, `mssql`, `mysql`, `oracle`, `postgres`, `snowflake`. E.g.
 
 ```shell
-uv pip install -U 'mostlyai[local, databricks, snowflake]'
+uv pip install -U 'mostlyai[local, snowflake]'
 ```
 
 ### Using Docker
@@ -276,7 +273,7 @@ As an alternative, you can also build a Docker image, which provides you with an
 
   <summary>Connect to the container</summary>
 
-  <p>You can now connect to the SDK running within the container by initializing the SDK in <code>CLIENT</code>> mode on the host machine.</p>
+  <p>You can now connect to the SDK running within the container from the host machine via an explicit <code>base_url</code>.</p>
 
   ```python
   from mostlyai.sdk import MostlyAI

diff --git a/docs/api_domain.md b/docs/api_domain.md
@@ -5,20 +5,11 @@ hide:
 
 # Schema References for `mostlyai.sdk.domain`
 
-This module is auto-generated to represent `pydantic`-based classes of the defined schema in the [Public API](https://github.com/mostly-ai/mostly-openapi/blob/main/public-api.yaml).
+This module is maintained manually to represent the SDK-supported `pydantic` domain classes.
 
 ::: mostlyai.sdk.domain
     options:
       show_root_heading: true
       show_root_full_path: true
       show_object_full_path: false
       show_root_toc_entry: false
-      filters:
-        - "!^Assistant.*"
-        - "!^Share.*"
-        - "!^ResourceShares"
-        - "!^LiteLlm.*"
-        - "!^DataLlm.*"
-        - "!.*PatchConfig.*"
-        - "!.*CloneConfig.*"
-        - "!^UsageReport.*"
diff --git a/docs/syntax.md b/docs/syntax.md
@@ -10,17 +10,12 @@ hide:
 ```python
 from mostlyai.sdk import MostlyAI
 
-# local mode (with TCP port)
-mostly = MostlyAI(
-    local=True,
-    local_dir='~/mostlyai',
-    local_port=8080,
-)
+# local mode (default)
+mostly = MostlyAI(local_dir='~/mostlyai')
 
-# client mode
+# client mode (explicit remote endpoint)
 mostly = MostlyAI(
-    base_url='https://app.mostly.ai', # or set env var `MOSTLY_BASE_URL`
-    api_key='INSERT_YOUR_API_KEY',    # or set env var `MOSTLY_API_KEY`
+    base_url='https://remote-sdk.example.com',
 )
 ```
 
@@ -164,67 +159,7 @@ c.open()
 c.delete()
 ```
 
-## Datasets
-
-Datasets can be used to train generators or create artifacts and may be created with or without a corresponding connector. Datasets are only available in `client` mode.
-
-```python
-# create a new dataset
-ds = mostly.datasets.create(config: dict | DatasetConfig)
-
-# retrieve a dataset by id
-ds = mostly.datasets.get(dataset_id: str)
-
-# list datasets (iterator with optional filters)
-for ds in mostly.datasets.list(
-    offset: int = 0,
-    limit: int | None = None,
-    search_term: str | None = None,
-    owner_id: str | list[str] | None = None,
-    visibility: str | list[str] | None = None,     # e.g., PUBLIC, PRIVATE, UNLISTED
-    created_from: str | None = None,               # YYYY-MM-DD
-    created_to: str | None = None,                 # YYYY-MM-DD
-    sort_by: str | list[str] | None = None,        # NO_OF_THREADS | NO_OF_LIKES | RECENCY
-):
-    print(ds.id, ds.name)
-
-# update an existing dataset (partial patch)
-ds = mostly.datasets._update(dataset_id: str, config: dict | DatasetPatchConfig)
-
-# delete a dataset
-mostly.datasets._delete(dataset_id: str)
-
-# download a file from a dataset
-content_bytes, filename = mostly.datasets._download_file(
-    dataset_id: str,
-    file_path: str,                                 # path inside the dataset
-)
-
-# upload a file to a dataset
-mostly.datasets._upload_file(
-    dataset_id: str,
-    file_path: str | Path,                          # local path to file
-)
-
-# delete a file from a dataset
-mostly.datasets._delete_file(
-    dataset_id: str,
-    file_path: str | Path,                          # path inside the dataset
-)
-```
-
 ## Miscellaneous
 
 ```python
-# fetch info on your user account
-mostly.me()
-
-# fetch info about the platform
-mostly.about()
-
-# list all available models
-mostly.models()
-
-# list all available computes
-mostly.computes()
 ```
diff --git a/docs/tutorials.md b/docs/tutorials.md
@@ -25,5 +25,3 @@ Explore synthetic data tutorials with the option to run them **either in Google
 | Close gaps in your data with **Smart Imputation**                      | [![Run on Colab](https://img.shields.io/badge/Open%20in-Colab-blue?logo=google-colab)](https://colab.research.google.com/github/mostly-ai/mostlyai/blob/main/docs/tutorials/smart-imputation/smart-imputation.ipynb)             | [View Notebook](./tutorials/smart-imputation/smart-imputation.ipynb)             |
 | Calculate accuracy and privacy metrics for **Quality Assurance**       | [![Run on Colab](https://img.shields.io/badge/Open%20in-Colab-blue?logo=google-colab)](https://colab.research.google.com/github/mostly-ai/mostlyai/blob/main/docs/tutorials/quality-assurance/quality-assurance.ipynb)           | [View Notebook](./tutorials/quality-assurance/quality-assurance.ipynb)           |
 | Enrich Sensitive Data with LLMs using Synthetic Replicas      | [![Run on Colab](https://img.shields.io/badge/Open%20in-Colab-blue?logo=google-colab)](https://colab.research.google.com/github/mostly-ai/mostlyai/blob/main/docs/tutorials/synthetic-enrich/synthetic-enrich.ipynb)           | [View Notebook](./tutorials/synthetic-enrich/synthetic-enrich.ipynb)           |
-| MOSTLY AI vs. SDV comparison: single-table scenario | [![Run on Colab](https://img.shields.io/badge/Open%20in-Colab-blue?logo=google-colab)](https://colab.research.google.com/github/mostly-ai/mostlyai/blob/main/docs/tutorials/sdv-comparison/single-table-scenario/single-table-scenario.ipynb)           | [View Notebook](./tutorials/sdv-comparison/single-table-scenario/single-table-scenario.ipynb)           |
-| MOSTLY AI vs. SDV comparison: sequential scenario | [![Run on Colab](https://img.shields.io/badge/Open%20in-Colab-blue?logo=google-colab)](https://colab.research.google.com/github/mostly-ai/mostlyai/blob/main/docs/tutorials/sdv-comparison/sequential-scenario/sequential-scenario.ipynb)           | [View Notebook](./tutorials/sdv-comparison/sequential-scenario/sequential-scenario.ipynb)           |
diff --git a/docs/tutorials/differential-privacy/differential-privacy.ipynb b/docs/tutorials/differential-privacy/differential-privacy.ipynb
@@ -6,17 +6,12 @@
    "metadata": {},
    "source": [
     "# Differentially Private Synthetic Data  <a href=\"https://colab.research.google.com/github/mostly-ai/mostlyai/blob/main/docs/tutorials/differential-privacy/differential-privacy.ipynb\" target=\"_blank\"><img src=\"https://img.shields.io/badge/Open%20in-Colab-blue?logo=google-colab\" alt=\"Run on Colab\"></a>\n",
-    "\n",
     "In this notebook, we demonstrate how a generator can be trained with differential privacy guarantees, and explore how the various settings can impact the data fidelity.\n",
-    "\n",
     "How Differential Privacy is applied:\n",
     "- Value ranges: DP is used to define value bounds for each column. The epsilon budget for this step is split evenly across columns.\n",
     "- Model training: DP-SGD by [Opacus](https://github.com/pytorch/opacus) is used for training, with a separate epsilon (and delta) value.\n",
     "- The total privacy budget is the sum of both parts.\n",
-    "\n",
-    "See also the schema reference for [DifferentialPrivacyConfig](https://mostly-ai.github.io/mostlyai/api_domain/#mostlyai.sdk.domain.DifferentialPrivacyConfig) for all available configuration parameters.\n",
-    "\n",
-    "For further background and analysis see also [this blog post](https://mostly.ai/blog/differentially-private-synthetic-data-with-mostly-ai) on \"_Differentially Private Synthetic Data with MOSTLY AI_\"."
+    "See also the schema reference for [DifferentialPrivacyConfig](https://mostly-ai.github.io/mostlyai/api_domain/#mostlyai.sdk.domain.DifferentialPrivacyConfig) for all available configuration parameters.\n"
    ]
   },
   {
@@ -244,7 +239,6 @@
    "metadata": {},
    "source": [
     "## Further exercises\n",
-    "\n",
     "In addition to walking through the above instructions, we suggest..\n",
     "* to experiment with different DP settings\n",
     "* to study the impact of the total size of the training data on final eps\n",
@@ -257,8 +251,7 @@
    "metadata": {},
    "source": [
     "## Conclusion\n",
-    "\n",
-    "This tutorial demonstrated how to train with and without differential privacy guarantees. Note: DP just provides additional mathematical guarantees for use cases that require these. However, given the other privacy mechanism in-built into the SDK, synthetic data can also without stricter DP guarantees be considered to be anonymous. See again [here](https://mostly.ai/blog/differentially-private-synthetic-data-with-mostly-ai) for a further discussion."
+    "This tutorial demonstrated how to train with and without differential privacy guarantees. Note: DP just provides additional mathematical guarantees for use cases that require these. However, given the other privacy mechanism in-built into the SDK, synthetic data can also without stricter DP guarantees be considered to be anonymous."
    ]
   }
  ],