Start with captions
Enable captions in a Microsoft Teams meeting. CapFlow reads the caption stream visible in your browser.
Built for the moments that matter in a meeting
CapFlow is a Chrome extension for Microsoft Teams captions. It brings live translation and meeting context into one clear workspace, so a fast conversation is easier to follow and revisit.
Early-stage product · Microsoft Teams captions required today
Let's align on the next release before we wrap up.
We should confirm the handoff and the owner today.
Bugün devir sürecini ve sorumlu kişiyi netleştirelim.
THE PROBLEM
When people speak across languages, even a short missed phrase can break the thread. Captions help, but switching between translation, notes, and the conversation adds friction.
Enable captions in a Microsoft Teams meeting. CapFlow reads the caption stream visible in your browser.
Read the original words alongside live translation, with the conversation kept in sequence.
Review meeting text and useful context after the conversation, without relying on memory alone.
THE PRODUCT
CapFlow is being built around a practical meeting workflow: capture what was said, understand it as it happens, and find it again later.
Talk to the teamKeep spoken content in a readable sequence during a Teams meeting.
See translated text together with its source caption, speaker, and place in the conversation.
Return to the transcript and extract the details you need after the call.
A browser extension and API-backed service designed for everyday use in multilingual meetings.
WHERE WE ARE HEADED
CapFlow starts with captions in Microsoft Teams. We plan to extend support to other online meeting platforms, then bring those conversations into a web app where people can capture, organize, search, and manage their meeting records in one place.
This is the product direction, not a description of features available today. Meeting capture and processing will be designed around user control and appropriate consent.
OUR AI ROADMAP
CapFlow currently uses another AI provider. We are planning a measured migration to Claude for translation, analysis, and meeting follow-up. Startup program credits would help fund that transition and its evaluation.
Evaluate Claude Haiku 4.5 for low-latency caption translation while preserving source context.
Evaluate the Claude Sonnet family for meeting analysis, structured outputs, and contextual chat.
Explore prompt caching for repeated context and the Batch API for post-meeting reports.
These are integration plans, not claims that Claude is already running in production. Model choices depend on latency, quality, privacy review, and cost testing.
THE TEAM
Rifat Arslan and the CapFlow team are working toward a more accessible way to follow and manage meetings. We are building the product step by step and are open to early feedback from people who face this problem every day.
Contact the teamLET'S TALK
Interested in early access, a pilot, or a conversation about CapFlow?