Anthropic is an American AI safety and research company founded in 2021 and headquartered in San Francisco, California. It develops the Claude family of large language models and is one of the world's leading frontier AI labs, focused on building reliable, interpretable and steerable AI systems.