wandb/weave
A Weights & Biases toolkit for tracing, evaluating, and debugging language model applications.
Not currently ranked — collecting fresh signals.
star history
Weave is a development toolkit for generative AI applications that provides observability and debugging capabilities for language model inputs, outputs, and execution traces. It enables developers to build rigorous evaluations for LLM use cases and organize information across the full AI development lifecycle from experimentation through production. The toolkit supports tracing for major LLM providers including OpenAI, Anthropic, Google AI Studio, and Hugging Face.
Frequently asked
- What is wandb/weave?
- A Weights & Biases toolkit for tracing, evaluating, and debugging language model applications.
- Is weave open source?
- Yes — wandb/weave is open source, released under the Apache-2.0 license.
- What language is weave written in?
- wandb/weave is primarily written in Python.
- How popular is weave?
- wandb/weave has 1.1k stars on GitHub.
- Where can I find weave?
- wandb/weave is on GitHub at https://github.com/wandb/weave.