---
id: tinker
version: "0.25.0"
license: Apache-2.0
license_treatment: permissive
maintenance: active
---

# tinker — The official Python SDK for the tinker API

License: permissive · Maintenance: active · Downloads: 1.2M/mo

## What it is and what it does

Tinker is the official Python SDK for accessing the Tinker API, maintained by Thinking Machines. It provides a client library to integrate Tinker's machine learning and data processing capabilities into Python applications. The package bundles a large dependency graph—including httpx for HTTP communication, pydantic for data validation, transformers for ML model support, and numpy for numerical operations—reflecting its role as a full-featured API client rather than a lightweight wrapper.

The SDK is actively maintained, with releases within days of the current date and a repository with 694 stars. It supports Python 3.9 through 3.13 (though installation requires 3.11+), runs on Windows, macOS, and Linux, and carries no known security vulnerabilities. The permissive Apache-2.0 license poses no restrictions on use or redistribution.

Use it for:

- Integrate Tinker's ML services into a Python application for model inference or data processing workflows
- Build data pipelines that call Tinker API endpoints from Python scripts or notebooks
- Prototype or deploy machine learning applications using Tinker's hosted models and APIs
- Automate batch processing or ETL tasks that depend on Tinker's computational services

## Worth the install?

AI-flagged interpretation of the facts above — verify before relying.

Tinker is the official Python SDK for the Tinker API, providing client libraries and utilities to interact with Tinker's machine learning and data processing services.

Yes. Tinker is the official SDK for its API, actively maintained with no security issues, low install friction, and a permissive license. Install it if you need to call the Tinker API from Python. The large dependency footprint (14 runtime packages) is typical for ML SDKs and reflects legitimate needs for HTTP, serialization, and model support rather than bloat.

## Install

pip install tinker
uv add tinker
poetry add tinker

## Installing tinker

Before you install: Low install friction with a pure-Python wheel. The package is actively maintained with a recent release and carries 14 runtime dependencies including httpx, pydantic, transformers, and numpy—a substantial but standard stack for ML-focused SDKs.

License in practice: Apache-2.0 is permissive; you can use, modify, and distribute this package freely in commercial and private projects, provided you include a copy of the license and note any changes.

Quickstart:

pip install tinker

import tinker
# Initialize client and call API methods as documented at tinker-docs.thinkingmachines.ai

Requires Python 3.11 or later; check your environment before installing.

Verify before relying:

- What specific Tinker API endpoints and features are exposed by this SDK version
- Whether transformers and numpy are always required or only for certain optional features
- Authentication method and credential management patterns for the Tinker API

## Package facts

- License: Apache-2.0 (permissive)
- Python support: supports_current
- Install friction: low
- Maintenance: active
- Downloads: 1.2M/month (top 5,000 on PyPI)
- Known vulnerabilities: none known

## Tags

tinker api python sdk, machine learning api client, tinker python library, api client for tinker, ml service integration python, tinker sdk, thinking machines api, api-client, machine-learning

[View on SkillFed](https://skillfed.io/packages/tinker) · [View on PyPI](https://pypi.org/project/tinker/)