Reflect

Reflect is built for moments where writing gets in the way. Meetings, lectures, conversations, or even something you’re watching.

Using a combination of speech-to-text systems like Whisper, Deepgram, AssemblyAI, and ElevenLabs, it captures spoken input with a focus on accuracy and speed.

That raw input is then processed using large language models, including Claude and GPT, to structure, summarize, and make sense of what was said. The goal is simple: turn unstructured audio into something clear and useful.

Designed with SwiftUI and backed by Realm, the experience focuses on clarity and ease of use. Capture first, understand later. Everything is organized so you can revisit, refine, and build on your thoughts over time.

Reflect sits between note-taking and thinking, helping you move from listening to understanding without friction.