Voice capture to searchable archive
The simplest and most powerful use case for SR-7: record freely, let the software handle the rest.
Daily voice journal
Record a quick debrief at the end of each day. Don’t worry about structure — just talk. SR-7 transcribes it, generates a title and summary, and adds it to your library. Over time, you build a searchable archive of what you were thinking, doing, and planning.
[Record 3 minutes about your day]
SR-7 generates:
Title: "Sprint retrospective thoughts and hiring update"
Summary: "Reflected on the sprint — delivery was solid but velocity
dropped due to onboarding. Noted two candidates to follow up with
after the panel interview on Thursday."Search your journal weeks later for “hiring” or “sprint” and find exactly what you said.
Meeting capture
Record meetings directly in SR-7, or import recordings from a hardware device afterward. The transcript becomes a searchable record, and the AI summary gives you the key points without re-listening.
Research and interviews
Record interviews, user research sessions, or field notes. Group them by project. Search across all recordings in a project to find patterns — every word is indexed once transcribed.
Working with AI agents via MCP
SR-7’s MCP server exposes your entire recording archive to AI tools. This turns voice into a data source that agents can read, search, and act on.
Search your recordings from an agent
You: "What did I say about the API redesign last week?"
Agent: [calls search with query "API redesign"]
Agent: "On March 28, you recorded a note titled 'API redesign direction'.
Summary: You decided to go with a REST-first approach, defer GraphQL,
and keep the existing auth layer. You mentioned wanting to discuss
rate limiting with Dan."Extract action items from recordings
You: "Go through today's recordings and pull out anything actionable"
Agent: [calls list_recordings — filters to today]
Agent: [calls get_recording for each — reads transcripts]
Agent: "I found 3 actionable items:
1. Send the revised proposal to the client (from your 2pm note)
2. Book a dentist appointment (from your morning memo)
3. Review the Q2 budget spreadsheet Lisa sent (from your post-meeting debrief)"Summarise a project
You: "Give me an overview of everything in the 'Product Launch' project"
Agent: [calls list_recordings with projectId]
Agent: [reads transcripts and summaries for all recordings]
Agent: "The project has 14 recordings over 3 weeks. Key themes:
- Messaging pivot: you moved from 'productivity' to 'clarity' positioning
- Launch date: targeting April 21, pending final QA
- Open question: pricing tier for teams — you keep going back and forth"Generate a written draft from voice notes
You: "I recorded some thoughts about our approach to on-device AI.
Turn those into a blog post draft."
Agent: [calls search with query "on-device AI"]
Agent: [reads transcripts from matching recordings]
Agent: [writes a draft based on your words and ideas]Your voice notes become first drafts. The agent preserves your thinking and phrasing while adding structure.
SR-7 + TR-1: voice to action
When you connect both SR-7 and TR-1 to the same AI agent, your voice recordings become a direct input to your task management system. Record your thoughts, and the agent turns them into structured, filed tasks.
Morning brain dump to task list
Record a stream-of-consciousness brain dump about everything on your mind. Then let the agent process it into your GTD system.
You: [Record 5 minutes in SR-7]
"I need to call the accountant about the Q1 filing... also the
design review is Thursday, I should prep slides for that... oh and
we're out of coffee filters... Maria asked about the API docs,
I should unblock her today... sometime this month I want to look
into that new database thing Dan mentioned..."
You: "Process my latest recording into TR-1"
Agent: [calls SR-7 search — finds the recording, reads transcript]
Agent: [calls TR-1 capture — "Call accountant about Q1 filing"]
Agent: [calls TR-1 create_task — "Prep slides for Thursday design review"
— dueDate: Wednesday, project: "Design Review"]
Agent: [calls TR-1 capture — "Buy coffee filters" — energy: low]
Agent: [calls TR-1 create_task — "Send Maria the API docs link"
— status: next, today: true]
Agent: [calls TR-1 create_task — "Look into new database Dan mentioned"
— status: someday]
Agent: "Created 5 tasks from your recording. 'Send Maria the API docs'
is flagged for today. 'Look into new database' is in Someday/Maybe."The agent applies GTD judgement — urgent items go to Today, vague ideas go to Someday/Maybe, and everything gets a clear title. You recorded a messy brain dump; you got a clean task list.
Post-meeting action extraction
You: [Record meeting in SR-7]
You: "Process the meeting I just recorded. Create tasks for anything
I committed to, and put follow-ups in Waiting For."
Agent: [calls SR-7 list_recordings — finds latest recording]
Agent: [calls SR-7 get_recording — reads transcript]
Agent: [calls TR-1 create_task — "Draft partnership proposal for Acme"
— project: "Partnerships", dueDate: next Friday]
Agent: [calls TR-1 create_task — "Waiting for: Jake to send pricing sheet"
— status: waiting]
Agent: [calls TR-1 create_task — "Schedule follow-up call with Acme
(after they review proposal)" — status: waiting]
Agent: "Created 1 action and 2 waiting-for items from the meeting."Weekly review with voice context
During your weekly review in TR-1, the agent can pull in context from your recordings to help you make decisions about stale tasks.
You: "Let's do the weekly review"
Agent: [calls TR-1 start_weekly_review]
Agent: [flags a stale task: "Research competitor pricing"]
Agent: "This task has been in Next Actions for 2 weeks. Want to
keep it, defer it, or drop it?"
You: "I think I talked about this — check my recordings"
Agent: [calls SR-7 search with query "competitor pricing"]
Agent: "You mentioned this on March 25: 'Competitor pricing isn't
urgent anymore — we decided to go with value-based pricing instead.
Maybe revisit in Q3.' Want me to move it to Someday/Maybe?"
You: "Yeah, defer it to July"
Agent: [calls TR-1 update_task — deferDate: July 1]Your recordings become the memory that makes reviews faster and decisions better-informed.
Markdown export for knowledge systems
SR-7 exports every recording as a Markdown file with YAML front matter — title, date, duration, transcript, and summary. These files integrate directly with Obsidian, Logseq, or any Markdown-based knowledge system.
Build a voice-first knowledge base
Enable auto-export in SR-7 settings. Every recording automatically becomes a Markdown file in your vault. Your voice notes sit alongside your written notes, fully linked and searchable.
Feed recordings into other tools
The exported Markdown files work as input for any tool that reads text — static site generators, search indexes, or custom scripts. The front matter makes them easy to filter and process programmatically.
Tips
- Record short and often. Five one-minute recordings are more useful than one rambling ten-minute recording. Shorter recordings get better titles and sharper summaries.
- Use projects to group context. An agent summarising a project with 20 recordings can give you a far better overview than one summarising your entire library.
- Star recordings you reference often. Stars give you a quick-access layer across projects — use them for key decisions, important ideas, or things you keep coming back to.
- Let transcription run automatically. Enable auto-transcribe on import so every recording is searchable the moment it lands in your library.
- Pair with TR-1 for a complete capture system. Voice in SR-7, tasks in TR-1, one agent connecting them. Nothing falls through the cracks.