This guide shows the end-to-end Late Meet workflow, from installation to meeting intelligence.
| Step | User Action | Late Meet Result |
|---|---|---|
| 1 | Install and build the extension | A Chrome-loadable dist/ folder is created |
| 2 | Load unpacked extension | Late Meet appears in chrome://extensions |
| 3 | Configure API keys | BYOK providers are ready for transcription and summaries |
| 4 | Join Google Meet | Late Meet detects the meeting context |
| 5 | Start Copilot | Audio capture and meeting intelligence begin |
| 6 | Open dashboard | Live summary, topics, decisions, and actions are shown |
| 7 | Review output | User can catch up, export, save, or discard meeting data |
- First-time users may see a guided onboarding wizard during setup.
- Keyboard shortcuts allow quick meeting actions without leaving the Meet tab.
- Timestamp-linked transcript navigation helps users jump from summary items to exact transcript chunks.
flowchart LR
subgraph Setup["Setup"]
build["Build extension"] --> load["Load dist/ in Chrome"]
load --> keys["Configure API keys"]
end
subgraph Meeting["Meeting flow"]
meet["Join Google Meet"] --> start["Start Copilot"]
start --> dashboard["Open side panel dashboard"]
end
subgraph Output["Meeting intelligence"]
review["Review summaries, decisions, and actions"]
end
keys --> meet
dashboard --> review
classDef setup fill:#eff6ff,stroke:#2563eb,color:#0f172a
classDef meeting fill:#fff7ed,stroke:#ea580c,color:#0f172a
classDef output fill:#f0fdf4,stroke:#16a34a,color:#0f172a
class build,load,keys setup
class meet,start,dashboard meeting
class review output
Build the project and load the generated dist/ folder in Chrome.
Use the popup as the lightweight entry point for checking extension state and opening configuration.
Add your own ElevenLabs and OpenAI keys in the options page.
When Late Meet detects a Meet session, the Start Copilot overlay appears inside the meeting interface.
The side panel gives a focused workspace for live meeting intelligence.
Dashboard areas include:
- Audio input status.
- Live summary.
- Current topic.
- Topics, decisions, action items, and people counters.
- Export summary action.
After Copilot starts, Late Meet captures tab audio through Chrome extension APIs and routes speech-to-text work through the configured provider. The dashboard should gradually reflect the active meeting state as transcript and intelligence updates become available.
flowchart LR
audio["Meet tab audio"] --> capture["Chrome tab capture"]
capture --> stt["Speech-to-text provider"]
stt --> transcript["Transcript updates"]
transcript --> intelligence["Summary, topics, decisions, actions"]
intelligence --> dashboard["Dashboard rendering"]
classDef source fill:#ecfeff,stroke:#0891b2,color:#0f172a
classDef process fill:#fff7ed,stroke:#ea580c,color:#0f172a
classDef output fill:#f0fdf4,stroke:#16a34a,color:#0f172a
class audio,capture source
class stt,transcript,intelligence process
class dashboard output
Late Meet is designed for users who join late or lose context during a call. The catch-up workflow should help the user understand the current meeting state without adding a bot participant or exposing private meeting data to a project-owned server.
Expected catch-up output includes:
- Current topic.
- Recent summary.
- Important decisions.
- Open action items.
- Participant or context signals when available.
The side panel dashboard is the main place to review generated meeting intelligence. Before exporting or sharing any output, users should check the summary, decisions, and action items for sensitive content.
Review the generated meeting intelligence before leaving the workflow:
- Live summary for the current meeting state.
- Topics for the main discussion areas.
- Decisions for confirmed outcomes.
- Action items for follow-up work.
- People counters and participant context when available.
Use the dashboard export action when you need a copy of the meeting summary outside the live session. Before sharing exported content, review it for private meeting details and remove anything that should not leave the team.
When the meeting is finished, decide whether the session output should be saved, exported, or discarded. This keeps the workflow aligned with Late Meet's local-first privacy model.
- Use the Storage Usage Dashboard to inspect saved sessions, see storage quotas, and clean up old meeting data.
- Screenshots show real extension states.
- Sensitive meeting names, codes, avatars, and account data are redacted.
- Documentation explains what the user should do and what they should expect to see.
- Setup, configuration, meeting usage, and dashboard review are all covered.