Feature
Meetings produce decisions, but those decisions are only valuable if someone captures and follows up on them. VidNotes analyzes meeting transcripts to automatically identify action items, decisions, owners, and deadlines. Instead of relying on manual note-taking during the meeting, you can focus on the conversation and let VidNotes extract the commitments afterward. The result is a clean, organized list of next steps that you can export and share with your team. For example, after a 30-minute sprint planning meeting where six tasks were assigned across four team members with varying deadlines, VidNotes will produce a structured list with each task, its owner, and the stated deadline, all linked to the timestamps where each commitment was made. The extraction uses OpenAI models through AIProxy with prompts specifically designed to identify commitment language, distinguish firm decisions from hypothetical discussions, and attribute tasks to the correct speakers.
Record your meeting and import it into VidNotes, or paste a link if it was recorded on a platform like Zoom, Teams, or Google Meet. The transcription captures everything that was said with timestamps, preserving the full context around each decision. You can import from your device, iCloud Drive, Google Drive, or Dropbox. For meetings recorded as video calls, the audio extraction pipeline handles the compressed audio codecs commonly used by meeting platforms.
After transcription, tap the action items option. The AIProcessingService reads the full transcript, identifies commitments, decisions, and assigned tasks using linguistic pattern recognition, and extracts them into a structured list. The AI distinguishes between firm commitments ('I will send the report by Friday') and hypothetical statements ('We might want to consider updating the docs'), ensuring only actionable items appear in the output. Each item includes a description, the mentioned owner, and a deadline when one was stated.
Review the extracted items, make any edits, and export the list as PDF or TXT. Share directly with your team through email, Slack, or your project management tool via the iOS share sheet. Each action item is linked to its source timestamp, so if there is any ambiguity about what was agreed, you or your team can tap through to hear the original conversation in context rather than debating what the notes meant.
Project managers
Get a clean list of action items after every meeting without relying on someone to take notes. Import directly into your project management workflow by exporting as TXT and pasting into Jira, Asana, or Linear. Use the timestamp links to resolve disputes about what was agreed when stakeholders remember the conversation differently.
Team leads
Share a concise summary of decisions and next steps with your team, including who owns each item and when it is due. Replace the post-meeting scramble to reconstruct what was said with a single tap that produces a shareable, accurate record of commitments made during the conversation.
Remote teams
Ensure alignment across time zones by sharing action item summaries from recorded meetings with team members who could not attend live. Remote team members get the same clarity as attendees, with the ability to tap into the original recording at any action item's timestamp to hear the full discussion context.
The action item extraction engine looks for linguistic patterns that signal commitments: phrases like 'I will', 'we need to', 'by Friday', 'let us schedule', and 'the decision is'. It understands context, so a hypothetical discussion about what someone might do is treated differently from a firm commitment. The AI also recognizes delegation patterns ('Can you handle the client follow-up?'), conditional commitments ('If the budget is approved, we will hire two more engineers'), and implicit deadlines ('before the next sprint' or 'by end of quarter'). This contextual understanding produces a focused list of genuine action items rather than a noisy dump of every sentence containing a future tense verb.
Each extracted action item is stored as a SwiftData ActionItem entity linked to the original VideoTranscript through a one-to-many relationship. This preserves the connection between the action item and the full context of the discussion, so you can always go back and listen to the exact moment a decision was made. The ActionItem entity stores the task description, owner attribution, deadline, and source timestamp. Because these are persistent local entities, your action item history accumulates across meetings, giving you a searchable record of commitments made over time.
Action items are generated in the language of the transcript using the AIProcessingService's language-aware prompt system. If your team meets in German, the extracted items will be in German with proper business terminology preserved. The language-specific prompts account for differences in how commitments are expressed across languages and cultures, such as the more indirect phrasing common in Japanese business meetings versus the explicit assignment patterns typical in English-language standups. This linguistic awareness reduces false negatives where commitments phrased in culturally specific ways might otherwise be missed.
No account required. Paste a video link and see it in action.