# Weco Documentation > Weco is production-grade autoresearch: point it at your code and an evaluation metric, and an automated research loop (AIDE) searches, tests, and ships measurable improvements as auditable, version-controlled code. Weco AI builds recursively self-improving AI. The `weco` CLI wraps the open-source AIDE agent with orchestration, steerability, and observability so teams can run the autoresearch loop in production — improving agent harnesses, prompts, ML models, and other performance-critical code against any metric you can measure. ## Getting Started - [Introduction](https://docs.weco.ai/): What Weco does and how the optimization loop works. - [Quickstart](https://docs.weco.ai/quickstart): Install and run your first optimization. - [How Weco works](https://docs.weco.ai/concepts/how-weco-works): The AIDE tree-search loop. - [Optimize with the CLI](https://docs.weco.ai/reference/optimize-with-cli): The `weco run` reference. - [Track any loop](https://docs.weco.ai/using-weco/track-any-loop): Observe external experiments in the Weco dashboard. - [Steerability](https://docs.weco.ai/using-weco/steerability): Guide the search with instructions and exemplars. ## Examples - [Agentic scaffold optimization](https://docs.weco.ai/examples/extract-line-plot) - [Prompt engineering](https://docs.weco.ai/examples/prompt-engineering) - [Model development](https://docs.weco.ai/examples/model-development) - [Fraud detection](https://docs.weco.ai/examples/fraud-detection) - [PyTorch optimization](https://docs.weco.ai/examples/pytorch-optimization) - [Triton kernel optimization](https://docs.weco.ai/examples/triton-optimization) - [CUDA kernel optimization](https://docs.weco.ai/examples/cuda-optimization) ## Resources - [FAQ](https://docs.weco.ai/resources/faq) - [Product site](https://www.weco.ai/): Positioning, use cases, and the examples gallery. - [weco CLI (GitHub)](https://github.com/WecoAI/weco-cli): Public mirror of the CLI. - [AIDE (open source)](https://github.com/WecoAI/aideml): The research agent behind Weco.