SigmaticOS is a lab-in-the-loop platform specializing in life sciences, announced by Sigmatic Sciences on March 3, 2026. It connects hypotheses derived from computational environments with the execution results from actual laboratories into a single, repeatable workflow. While typical research software handles only one segment, such as model execution, data analysis, or equipment control, SigmaticOS aims for an operational layer that orchestrates both in silico workflows and wet-lab automation, while also tracking the process. According to the official announcement, it utilizes over 100 agents and integrates machine learning models trained on customer data, along with large language models (LLMs), recursive language models (RLMs), and retrieval-augmented generation (RAG). In essence, it's like moving the task instructions and result reports that researchers previously manually transferred between the computational and laboratory environments into a digital control center shared by multiple specialized agents. In existing computational biology pipelines, researchers often have to re-enter the candidates recommended by the model into a separate experiment management system, organize the results from the equipment, and then reprocess them as model training data. At each boundary where tools and personnel change, experimental conditions, data lineage, causes of failure, or the context of decision-making can be lost. The key difference of SigmaticOS is that it aims for a closed-loop research flow that doesn't just end with reporting computational results but feeds them back into actual experiment execution and subsequent learning. It focuses on a structure where multiple agents share roles, rather than a single generative AI assistant, and treats computational research and laboratory operations as a single system that continuously exchanges feedback, rather than separate automation areas. In the exploration of new drug candidates, a scenario can be configured where researchers calculate candidate priorities based on internal data, connect subsequent wet-lab validation through an agent workflow, and then feed the obtained results back into the machine learning model's training process. In cell-based assays, in silico condition design, experiment execution requests, result tracking, and the suggestion of the next experiment conditions can be combined into a single iterative cycle. In process or analytical method optimization, it is possible to reference previous experiment records using RAG and apply LLM- and RLM-based inference to coordinate the next steps. However, supported equipment, data formats, agent-specific permissions, validated throughput, model performance metrics, and the scope of regulatory compliance could not be confirmed based solely on the provided information; therefore, official technical documentation and supplier validation are necessary before actual implementation.

SigmaticOS is a lab-in-the-loop platform specializing in life sciences, announced by Sigmatic Sciences on March 3, 2026. It connects hypotheses derived from computational environments with the execution results of actual laboratory experiments into a single, repeatable workflow. While typical research software handles only one segment, such as model execution, data analysis, or equipment control, SigmaticOS aims for an operational layer that orchestrates both in silico workflows and wet-lab automation, while also tracking the process. According to the official announcement, it utilizes over 100 agents.