personal memory agent
speaker insight tweaks
+14
-29
+1
-2
think/insights/speakers.json
+1
-2
think/insights/speakers.json
···
1
1
{
2
2
"title": "Meeting Speakers",
3
-
"description": "Detects video meetings in the segment and extracts participant names from screen and conversation.",
3
+
"description": "Detects meetings in the segment and extracts participant names from screen and conversation.",
4
4
"frequency": "segment",
5
5
"occurrences": "For each meeting detected, extract start/end times and the list of participant names. If segment spans two meetings, create separate occurrences.",
6
-
"contains": "Meeting detection with time bounds and participant name lists extracted from visible participant panels and conversation.",
7
6
"color": "#ff5722"
8
7
}
+13
-27
think/insights/speakers.txt
+13
-27
think/insights/speakers.txt
···
2
2
3
3
## Objective
4
4
5
-
Identify meetings in this segment and extract participant names. This output feeds into occurrence extraction to populate meeting participant lists.
5
+
Identify meetings and multi-person discussions in this segment and extract possible participant names.
6
6
7
7
## Input Types
8
8
9
-
### With Screen Activity (Live Capture)
9
+
### With Screen Activity
10
10
11
11
A meeting is detected when screen activity shows a video conferencing application:
12
12
- Zoom, Google Meet, Microsoft Teams, Webex, or similar
···
14
14
- Meeting controls or UI elements present
15
15
16
16
Extract participant names from:
17
-
1. **Visible participant list/panel** on screen (highest confidence)
17
+
1. **Visible participant list/panel** on screen
18
18
2. **Names spoken in conversation** - direct address ("Thanks, Sarah"), mentions ("John was saying...")
19
19
3. **Self-introductions** ("Hi, I'm Alex from...")
20
20
21
-
### Audio Only (Import)
21
+
### Audio Only
22
22
23
23
When no screen activity is present, a meeting is detected when audio shows:
24
24
- Multiple speakers with conversational turn-taking
25
25
- Meeting-style patterns: agenda items, discussion, decisions, action items
26
26
- Professional/work context evident from topics
27
+
- Personal discussions amongst friends or family
27
28
28
29
Extract names from:
29
30
1. **Names spoken in conversation** - direct address and mentions
30
31
2. **Self-introductions**
31
-
3. **Setting/topics metadata** if available from transcription
32
+
3. **Any setting metadata** if available from transcription
33
+
4. **Facet** if given, possible entities based on the given facet
32
34
33
35
## Boundary Detection
34
36
35
-
If the segment spans two distinct meetings (one ending, another starting):
37
+
If the segment spans two distinct meetings with different participants (one ending, another starting):
36
38
- Look for farewell patterns followed by new greetings
37
39
- Different participant set appears
38
40
- Significant topic shift with new introductions
···
41
43
42
44
## Output Format
43
45
44
-
Structure your output clearly for occurrence extraction:
45
-
46
-
### Meeting Detection
47
-
48
-
State whether meeting(s) were detected and the detection basis:
49
-
- **Detected**: Yes / No
50
-
- **Source**: Screen (video call UI visible) / Audio-only (conversation patterns)
51
-
52
-
### Meeting 1
46
+
Structure your output as clean well formatted markdown with the following information included:
53
47
54
48
- **Time**: HH:MM:SS - HH:MM:SS
55
49
- **Platform**: Zoom / Meet / Teams / etc. (if visible, omit for audio-only)
56
50
- **Participants**:
57
-
- Name (evidence: where/how detected)
58
-
- Name (evidence: where/how detected)
59
-
60
-
### Meeting 2 (if segment contains two meetings)
51
+
- Speaker Name
52
+
- Another Speaker Name
61
53
62
-
Same structure as above.
54
+
Your result may have 0, 1, or 2 of the above sections. Our goal is to identify all the possible speakers based on the evidence.
63
55
64
56
## What to Ignore
65
57
66
58
- Podcasts, videos, or streaming content being watched
67
59
- Music or background audio
68
-
- Single-speaker recordings without meeting context
69
-
- Phone calls (no video conferencing UI)
70
60
71
61
## No Meeting Case
72
62
73
-
If no meeting is detected, output:
74
-
75
-
### Meeting Detection
76
-
- **Detected**: No
77
-
- **Reason**: [Brief explanation - e.g., "No video call UI visible, single speaker only"]
63
+
If no meeting is detected, output just "No meetings detected"