SELF / Review round 02

Your evening prepares tomorrow

A practical checklist for you and the voice agent. Reflect on reality, plan deliberately, prepare your environment, then let the closing process handle the records and audios.

Reviewable design · automations not activated

The evening has three parts

1Reflect

Read the digest together. Add the context the computer cannot know.

Agent writes opening notes

2Plan and track

You write the plan, calendar and tracker while talking. The agent listens and helps you think.

You write manually

3Prepare and finish

Get tomorrow physically ready. Confirm each item, then close the session.

Agent checks completion
The writing boundary: after the opening reflection, the agent stops live note-taking. It keeps listening, coaching and tracking checklist state. Full conversation capture continues. Closeout saves that transcript without overwriting what you manually wrote.

At 17:30, the future Daily Digest process prepares a pinned Codex task. At around 18:00, use that same task and invoke plan-tomorrow-and-reflect. It opens the digest, dashboard, next-day plan, tracking Sheet and this checklist.

These triggers are the reviewed design. No new scheduler or skill has been activated by this page.

The shared completion checklist

The agent uses a goal checklist, waits for verbal confirmation and does not confuse talking about something with doing it. Sickness, travel and deliberate changes are recorded explicitly. Checkboxes here are a usable preview, not live Notion state.

After closing: gym and sauna reminders, then your chosen bedtime. Their later completion stays pending until you report it. The session does not remain open waiting for tomorrow.

Your next 24 hours

17:30 · The evidence is prepared

Computer/Codex activity, Missive conversations and WisprFlow meetings are compared with the saved plan and calendar. Tomorrow’s priorities are only suggestions.

18:00 · Reflect → plan → prepare

Use the checklist above in the prepared Codex task. Finalize before gym and sauna. Both audio guides are generated from the plan you actually saved.

Wake · Audio 1

AirPods in, press play. Brief mindset coaching and the first task’s exact start. Begin working within about ten minutes. Four repeats of the fixed 30-minute track support the first two-hour block.

After the first block · Audio 2

Run and sunlight, mindset coaching, cold shower, next-task briefing, guided meditation and the next focus block. Keep the remaining plan realistic and use normal pause/resume when the day intervenes.

Rest of day · Follow the actual plan

NSDR and another focus block when planned. Calls, communications, food, walks and leisure have space. Record actual work times, then reflect again that evening.

Tracking stays manual

Use the new copy of your existing tracking tab. You enter the data; the agent listens. Original history stays intact. Use row 107 for the complete six-block work-hours total; the preserved legacy total in row 37 covers only four blocks.

Open tracking SheetOpen dashboard

Time and completion

Actual sleep/wake; focus starts/ends and totals; Wake to Flow; each audio; planning start/finish; time-blocking and first-task preparation.

Recovery and preparation

Run, cold shower, meditation, NSDR, walks, movement, food, shakes, water, clothes and desk. Caffeine after 13:30 and daily wins follow your screenshot.

Business facts

Cold emails, positive replies, sales calls and Wise balance. Read CRM/Instantly yourself. Blank or unknown does not become zero.

Two play buttons, one saved plan

Audio 1 · Wake to Flow

Personalized short introduction → first actions → two-hour focus music.

From tomorrow’s final plan, why the main task matters and your approved coaching preferences.

Audio 2 · Run, reset and continue

Personalized coaching → cold shower → task brief → 3 + 7 minute meditation → two-hour focus → supplied NSDR → next planned focus block.

From that same plan, the full evening conversation, personal history and relevant mentor material.

Both arrive in Telegram with segment timestamps. The supplied NSDR is 10:42, so its full guidance is preserved. No prerecorded track can detect that you have reached the desk: pause it when needed.

See exactly where the audio content comes from →

What changed after your review

Added: dedicated Daily Digests database, human email evidence, calendar comparison, same-task planning, opening-only live notes, explicit physical checklist, populated examples and a copied tracking tab.

Retained: manual planning/tracking, both long audio structures, daily reflection, personal memory retrieval and Telegram delivery as the future workflow.

Deferred: installing the four runtime skills, scheduling the digest, live ingestion/indexing and actual audio generation. This review makes their inputs, behavior and outputs precise.