Comet is an AI developer platform used by machine learning and AI engineering teams to track experiments, evaluate models, and observe LLM and agent behavior in production. Founded in 2017 by Gideon Mendels and Nimrod Shabtay, the New York-based company built its early reputation as an experiment tracking tool for traditional machine learning, competing with Weights and Biases and MLflow.
As generative AI workloads grew, Comet expanded into LLM observability with Opik, an open-source platform that lets developers log traces, run automated evaluations, and monitor production agents. Opik supports more than forty frameworks and providers, including OpenAI, Anthropic, LangChain, LlamaIndex, and Hugging Face, and has accumulated tens of thousands of GitHub stars since its 2024 launch. The platform connects tracing and evaluation results directly to code-level remediation suggestions.
Comet has raised roughly $70 million across multiple rounds, including a Series B led by OpenView Venture Partners, with additional investors such as Scale Venture Partners and Trilogy Equity Partners. Customers include Uber, Netflix, Etsy, NatWest, Autodesk, and Stellantis, and the company reports usage by more than 150,000 developers across 10,000 teams worldwide.
Comet's positioning combines a paid enterprise platform with a permissively-licensed open source core, giving teams a low-friction entry point that can graduate into governed enterprise deployments. The company competes with LangSmith, Arize, Weights and Biases, and Helicone in the LLM observability space.