Kurt.ai

Meetings should drive action, not create work. Kurt.ai joins calls, understands decisions, and instantly creates tasks in Linear — making calls productive and tasks complete.

Comment

Team

Bartek Pichalski Wojciech Mandrysz Olena Nikolenko

Inspiration

Meetings should be where work gets done, not where it gets stuck. But too often, they create chaos—scattered action items, forgotten decisions, and endless manual follow-ups. Product teams waste hours tracking tasks instead of building.

Our story starts with two personas: developers frustrated with empty task descriptions that leave them guessing about what to do, and PMs drowning in the manual effort of updating those tasks and tracking progress. Both sides were spending too much time managing the details instead of pushing projects forward.

We saw a better way. A world where decisions don’t disappear into Slack threads. A world where meetings don’t generate more work—they generate progress.

So we built Kurt.ai. Not another note-taking tool. Not another assistant. But an AI agent that takes action.

What it does

Kurt.ai joins your Google Meet calls, listens like a real team member, and manages ticket creation.

  • Identifies actionable tasks and creates detailed Linear tickets, capturing the full context, priorities, deadlines and ownership of each decision.
  • Updates tasks with new information and reassigns them as priorities shift.
  • Organizes project requirements in real time, ensuring teams stay aligned and informed throughout the meeting.

How we built it

To make this vision a reality, we combined:

  • Real-time speech-to-text processing to capture discussions.
  • A WebRTC integration.
  • Automated task creation via Linear SDK.
  • Extension for Google Meets.

Accomplishments that we're proud of

  • Achieved high accuracy in identifying tasks and assigning ownership with minimal false positives.

What's next for Kurt.ai

  • ElevenLabs integration enabling Kurt to speak and signal actions in real-time.
  • Expanding integrations to other task-managers beyond Linear to support Jira, Asana, and Notion.
  • Adding the different types of actions like real-time knowledge extraction from Confluence or other knowledge bases to highlight past decisions, key data, or unresolved topics during the call.
  • Audio recording attached with the task description, allowing the developer to listen to the context behind the task.
  • Slack integration

...and many more.

Built With

Share this project:

Updates