Clusy
Clusy is an agent-native notebook platform that automates end-to-end machine learning and data science workflows. Users describe desired outcomes in natural language, and an AI agent plans the work, writes and executes notebook cells on cloud CPUs or GPUs, and delivers results that remain fully inspectable, editable, and re-runnable. The platform keeps the transparency of a traditional notebook while handling infrastructure, data sourcing, and model selection. Unlike empty notebook environments, Clusy’s agent actively participates in the workflow. It sources datasets from public repositories or user connections, inspects data quality, selects appropriate model architectures and compute resources, and executes training or analysis. The notebook interface supports branching and parallel experiments, allowing users to compare different approaches side by side. Clusy offers tiered plans with a free CPU sandbox and paid options that unlock access to GPU sandboxes (T4, L4, A10, L40S, A100, H100, H200) and advanced AI models (DeepSeek, Kimi, Claude, GPT). Users can connect their own data via uploads, Hugging Face, or direct integrations with Databricks and Snowflake. The platform is used for tasks like fine-tuning large language models, reproducing research models, automating data preprocessing, and conducting market research. Testimonials highlight its ability to handle complete workflows from a single prompt, saving time and reducing manual overhead for ML engineers and researchers.
- Category
- AI Agent
- Pricing
- Unknown
Problem solved
Traditional machine learning workflows require manual data sourcing, preprocessing, model selection, and infrastructure management, often slowing down iteration. Clusy automates these steps by using an AI agent that interprets natural language prompts to execute end-to-end ML tasks in a cloud notebook environment, giving users code-level control and reducing time to insight.
Use cases
- Fine-tuning large language models with a single prompt
- Reproducing machine learning models from research papers
- Automated dataset discovery and preprocessing
- Idea validation and market research via data analysis
- Post-training open-weight models for specialized tasks
- Parallel experiment branching and comparison
- Connecting to live data warehouses (Databricks, Snowflake) for analysis
Audience
- Developers
- ML Engineers
- Data Scientists
- Researchers
- Business Users
Pros
- Natural language driven automation of end-to-end ML workflows
- Transparent notebook interface preserves code control and editability
- Built-in cloud GPU range from T4 to H200 for scaling compute
- Supports experiment branching and parallel runs for rapid iteration
- Integrates with external data sources like Databricks and Snowflake
Cons
- Free plan limited to CPU, entry GPU requires paid plan
- Proprietary platform with no self-hosted option
- Potential vendor lock-in for complex pipelines
- High-end GPU costs can accumulate quickly for heavy usage
- Relatively new, community and third-party integrations may be limited
Official website: https://www.clusy.io/
Canonical LambdaBase page: https://www.lambdabase.com/tools/clusy