FDA clears first software as a medical device with a patient-facing large language model

UpDoc announced June 25, 2026, that it received FDA 510(k) clearance for UpDoc V1.0, the first software as a medical device built on a patient-facing large language model. The device, cleared December 23, 2025, supports insulin and medication management for adults with type 2 diabetes under licensed clinician supervision, officials said.

The software provides patient-facing instructions within a clinician-defined care plan, enabling medication titration and related care coordination tasks, sources said. Unlike prior AI-cleared devices that primarily analyze medical images or flag anomalies, UpDoc V1.0 incorporates a large language model (LLM) directly in a patient-facing workflow, marking a regulatory first in this category.

The UpDoc V1.0 device, cleared under FDA 510(k) number K253281 on December 23, 2025, is designed to support insulin and medication management for adults with type 2 diabetes, operating under licensed clinician supervision, according to company officials and public FDA records.

UpDoc framed the clearance as a significant milestone for patient-facing LLMs in regulated medical software, with the device’s scope narrowly defined to specific clinical tasks consistent with the FDA’s substantial equivalence review process under the 510(k) pathway, company statements and regulatory analysis show. The predicate device cited for this clearance was the non-LLM d-Nav system, which also supports diabetes medication management, sources confirmed. The company emphasized that UpDoc V1.0 does not independently make treatment decisions but functions within protocol-bounded, clinician-supervised parameters.

The FDA’s clearance did not require a traditional clinical trial, according to reporting on the submission process. Instead, the company provided design controls, clinical evidence, and a Predetermined Change Control Plan (PCCP) to address the device’s total product life cycle, as outlined in FDA guidance documents on AI and machine learning-enabled medical devices. The FDA has indicated it is exploring ways to identify and tag devices incorporating foundation models, including LLMs, in its regulatory framework, according to agency materials.

UpDoc announced the clearance alongside news of an $18 million seed funding round, which included participation from Mayo Clinic and Eli Lilly, industry sources reported. The company is positioning itself as an enterprise-ready clinical AI firm, focusing on real-time patient care and care coordination, with initial deployment discussions reportedly targeted at major health systems. This strategic approach aligns with the device’s narrowly scoped clinical application and supervised use model.

Regulatory experts note that the classification of software as a medical device depends heavily on the software’s intended use, particularly whether it provides specific diagnostic or treatment directives. UpDoc’s platform, by operating under licensed clinician oversight and within defined protocols, fits within existing FDA pathways for SaMD, according to regulatory commentary. This approach contrasts with open-ended conversational assistants, which pose greater regulatory challenges due to potential risks associated with autonomous clinical decision-making.

Most AI/ML medical devices cleared by the FDA to date have been authorized through the 510(k) pathway rather than De Novo or premarket approval (PMA), with estimates suggesting that approximately 97% of cleared AI/ML devices as of late 2023 used 510(k) clearance, according to secondary analyses cited in industry reporting. The UpDoc device’s clearance exemplifies how tightly bounded, clinician-supervised AI tools can meet regulatory requirements without extensive clinical trials.

Academic research highlights that LLMs can produce outputs resembling medical device functions even when constrained to decision-support roles, underscoring the FDA’s emphasis on transparency, human oversight, and safety in regulating patient-facing AI tools. The agency’s AI-enabled medical device webpage states ongoing efforts to refine identification and oversight methods for devices incorporating foundation models.

UpDoc’s clearance represents a step in the evolving regulatory landscape for AI-enabled medical devices, with the FDA continuing to develop policies that address the unique challenges posed by foundation models and large language models in healthcare settings.