1029.2MtokensCaptured live
683builds
3251.7hbuild time
141.7klines
15dactive 30d
6wweek streak
Real coding sessions · never self-reported
Highlights
Where the tokens go
50% building · 1027.9M classified tokens
Building511.0M50%
Verifying203.3M20%
Thinking157.4M15%
Delegating156.2M15%
32 builds predate classification and aren't included.
6
week streak
35.9M
token record
1mo ago
215
prompt record
2mo ago
2d ago
Fixed weather indicator calculation bugs in brief generation
meteo-brief
Top 16% session this week
950 sessions tracked this week across all builders
877.7ktokens
11prompts
56:48time
99lines
Claude Codeclaude-sonnet-5$16.71
mostly verifying
Fixed precision bugs in weather hazard and precipitation calculations; all 378 tests passing after merging both fixes.
3d ago
Merged weather hazard features and added validation checks
meteo-brief
MARATHON
25.5x your typical session
4.9M tokens vs typical 192.4K
4.9Mtokens
28prompts
7:07:16time
43lines
Claude Codeclaude-sonnet-5Markdown$62.60
mostly delegating
Merged three PRs and added cross-tier confidence validation checks after full test suite confirmation.
5d ago
Fixed judge logic marking valid entries as failed
meteo-brief
FULL SEND
19.5x your typical session
3.7M tokens vs typical 192.3K
3.7Mtokens
10prompts
4:45:35time
14lines
Claude Codeclaude-opus-5$85.34
mostly verifying
Fixed validation judge that was incorrectly invalidating valid weather data entries by contradicting its own rules.
5d ago
Resolved promoted carousel review feedback
Esdeveniments
FULL SEND
Longest session this week
52h 10m
1.7Mtokens
10prompts
52:10:26time
322lines
Claude Codeclaude-sonnet-5TypeScript$15.18
mostly building
Fixed an event filtering bug and added test coverage for the promoted carousel.
5d ago
Fixed development server by resyncing stale dependencies
Esdeveniments
Top 20% build time this week
2071 sessions tracked this week across all builders
253.9ktokens
3prompts
6:11time
Claude Codeclaude-opus-5$4.01
mostly verifying
Restored dev server functionality after resyncing node_modules against yarn.lock.
5d ago
Fixed hero rendering on slow networks and mobile devices
Repos
FULL SEND
14.1x your typical session
2.7M tokens vs typical 191.9K
2.7Mtokens
18prompts
7:04:28time
1,395lines
Claude Codeclaude-opus-5TypeScript$99.66
mostly building
Fixed how the hero component adapts to slow network connections and mobile viewport sizes.
6d ago
Built a voice profile analyzer from speech transcripts
Repos
FULL SEND
Most lines changed this week
Previous best: 1,530 lines
1.6Mtokens
17prompts
3:34:40time
2,283lines
Claude Codeclaude-sonnet-5Markdown$31.82
mostly building
Built voice analysis tools that learn from your actual speech and validate communication consistency.
6d ago
Fixed iOS Safari video playback after debugging
eltempsavui
LOCKED IN
Top 20% session this week
936 sessions tracked this week across all builders
530.8ktokens
2prompts
12:01:28time
91lines
Claude Codeclaude-opus-5JavaScript$12.03
mostly building
Fixed iOS video playback, deployed to preview for device testing.
6d ago
Fixed video playback and scroll animation timing
eltempsavui
BUG HUNT
Biggest session this week
6.7M tokens
6.7Mtokens
22prompts
9:02:34time
628lines
Claude Codeclaude-sonnet-5TypeScript$88.86
mostly verifying
Debugged and fixed multiple video and animation timing issues.
7d ago
Investigated video generation issues across frames to identify root causes
eltempsavui
MARATHON
Longest session this week
15h 47m
3.5Mtokens
34prompts
18:54:29time
377lines
Claude Codeclaude-sonnet-5TypeScript$54.97
balanced mix
Traced stochastic failures in video diffusion model output to content-dependent issues.
8d ago
Investigated API quota errors and identified root cause
meteo-brief
Top 14% build time this week
1992 sessions tracked this week across all builders
258.7ktokens
2prompts
15:33time
Claude Codeclaude-sonnet-5$1.45
mostly verifying
Debugged API quota issues affecting the service and confirmed meteo-brief's error handling is working correctly; root cause originates in a different process.
8d ago
Debugged meteo-brief
Esdeveniments
DEEP WORK
19.0x your typical session
3.6M tokens vs typical 188.1K
3.6Mtokens
8prompts
12:17:07time
271lines
Claude Codeclaude-sonnet-5TypeScript$26.78
mostly building
7m session running commands and testing. 75K tokens across 3 prompts using claude-sonnet-5.