Our take
Inactiveno commits in over a year- Activity
- 20
- Community
- 43
- Issues
- 21
- Pull requests
- 21
+2 stars in the last 8 days
Curie offers a concrete research-automation workflow, but its inactive status and lack of recent maintenance make it a cautious choice for new or production-critical work.
Good fit if
- You are a researcher or ML practitioner exploring AI-run experiments and can review the generated code and findings yourself.
- You want to provide your own dataset or starter code and are comfortable using a Python package with an external model API.
- You can evaluate the project as-is under Apache-2.0 terms without depending on ongoing maintenance.
Look elsewhere if
- Your work requires recent fixes, active issue handling, or a current stable release. The repository has had no commits or closed issues in over a year, and no releases in over a year.
- You need a mature, actively maintained framework for critical research workflows. The project is marked inactive, and its latest release is a prerelease.
- Your team needs a tool that does not rely on third-party model APIs or cannot accommodate the Docker setup described in the README.
All health signals
| Last commit | 2025-09-28 (1 year ago) |
|---|---|
| Commits, last 90 days | 0 |
| Releases, last 12 months | 0 (latest v0.1.4, 2025-06-03, pre-release) |
| Contributors | 6 (top contributor: 60% of commits) |
| Issues closed, last 90 days | 0 |
| Pull requests merged, last 90 days | 0 |
| Project age | 20 months |
Checked on 2026-10-09 with the GitHub API.
Overview
Curie is a Python framework for automating scientific experimentation with AI agents. It aims to take a research question through hypothesis formulation, experiment implementation and execution, result analysis, and reporting, with verification steps intended to support rigor and reproducibility.
It is aimed at researchers and machine-learning practitioners who want an agent to investigate a question using their own dataset or starter code. Curie is more than a question-answering tool: it runs experiments and produces artifacts that can be reviewed and used to reproduce the work.
Key Features
- Turns a research question into an experiment workflow, from hypothesis formation through result interpretation.
- Runs generated experiment code and records the process in a workspace.
- Provides verification modules intended to improve procedural reliability and reproducibility.
- Accepts user-provided datasets and optional starter code and requirements files.
- Supports machine-learning tasks as well as examples in system analysis and scientific discovery.
- Produces experiment reports and result files for reviewing findings.
- Offers an AutoML feature for searching for machine-learning approaches on a task.
Use Cases
- A researcher with a focused hypothesis can use Curie to compare approaches, such as sorting algorithms across input distributions, and inspect the generated report and results.
- A data scientist can test model robustness on a supplied dataset, for example comparing classifiers under added noise.
- A machine-learning practitioner with a dataset and existing code can ask Curie to explore ways to improve model performance.
- A team evaluating AI research agents can examine Curie’s experiment workflow and related EXP-Bench materials to study automated research capabilities.
What you need
Detected in the repository
- Python (from setup.py)
- Automated checks on GitHub Actions
License in plain words
Apache-2.0permissive
- Commercial use: yes
- Modify and redistribute: yes
- You must keep: the license, the NOTICE file and a note of your changes
- Share your changes: no
- Includes an explicit patent grant from the contributors:
A summary, not legal advice: the LICENSE file is what applies.
Getting Started
The README recommends installing Docker before running Curie. Install the Python package with:
pip install curie-ai
See the README for API-key setup, examples, and manual installation instructions.
Alternatives
- Biomni: Biomni targets biomedical research tasks with domain tools and data, while Curie focuses on planning and running scientific experiments more broadly.
| Project | Language | License | Stars | Status |
|---|---|---|---|---|
| Curie | Python | Apache-2.0 | 370 | Inactive |
| Biomni | Python | Apache-2.0 | 3.9k | Quiet |
Considerations
- Repository activity is inactive: there have been no commits in over a year, no releases in over a year, and no issues closed or pull requests merged in the last 90 days. The project has six contributors, and contribution activity is not reported as concentrated in one person.
- The latest listed release is a prerelease, so teams should verify whether its current behavior and compatibility meet their needs.
- Curie requires Python and uses model-provider API keys in the README examples. The README also lists Docker as a prerequisite, so confirm the deployment requirements for your intended setup.
- The README recommends a GPU for ML training tasks, but does not make one a general prerequisite.
- The Apache-2.0 license permits broad use subject to its license terms; review those terms for your distribution and compliance needs.
Found this useful?
Share it with someone who would like Curie.