Weekly Reset
--d --h --m
FRI 06:00 GT
Pacing Control Center
No local actions recorded yet. Save a plan or paste /usage to start history.
Weekly Consumption · Live
80% / 100%
You are 9% over pace with 1d 14h remaining.
PACE
20%
Remaining
1.4%/hr
Required Pace
4.2%/hr
Current Burn
Burn Rate · 60min avg
4.2%/hr
▲ 200% vs required
— — — TARGET 1.4%/hr
⚠ EXHAUSTION PROJECTED: 10:42 PM THU
Surface Attribution
Claude.ai
49%
Code
25%
Cowork
6%
Excel
20%
API
0%
Chat is your largest spend. Two conversations >80K tokens drove 31% of total this week.
! RECOMMENDATION
!
Switch to Sonnet for the next 4 hours, or budget exhausts at 10:42 PM Thursday — 7 hours before reset.
Confidence 91% · based on 14 similar pace deviations · predicted savings 8.4%
Burn vs. Pace
Cumulative consumption · this cycle
Daily Distribution
Pattern-weighted · 8-week avg
▬ Actual| Target▬ Over
Model Mix · This Week
OPUS
55% ~$2.20
SONNET
35% ~$0.42
HAIKU
10% ~$0.04
Burn multiplier vs. all-Haiku 3.1×
Circuit Breaker
ARMED · MONITORING
Auto-alerts fire at 50%, 70%, 90%. Next: 90%.
50% alert✓ Fired Tue 09:14
70% alert✓ Fired Tue 22:47
90% alert— Pending
100% lockout— Pending
Hidden Cost Audit
Project KB overhead~32K/turn
Conversation history~18K avg
MCP schema injection~4K/call
Memory summaries~2K/session
Total invisible overhead ~56K tokens
Usage History
7-day pattern · week-over-week · trend detection
THIS WEEK
80%
Total Used
▲ 12% vs last week
23%
Peak Day
Monday
74%
Pacing Efficiency
▲ 3% vs last week
0
Lockouts
No lockouts this week
11.4%
Avg / Day
Target 14.3%
Daily Burn — Week over Week
Bars = actual · Line = 8-week average · Dashed = target
Day Breakdown
This week vs last week vs 8-week avg
Day This Week Last Week 8W Avg vs Target Pattern
Pattern Detection Engine
Optimize Engine
Actionable recommendations · ranked by savings impact
22%
Potential Savings
If all recs applied
6
Active Recommendations
0 applied this week
HIGH
Priority Level
Budget at risk
6h
Time to Safe Zone
If top rec applied now
Model Routing Guide — When to Use What
Yield Analysis
ROI per conversation · waste detection · value scoring
⬆
Import CodeBurn Data
Drop a CodeBurn JSON export here, or click IMPORT above.
Don't have CodeBurn? Click LOAD DEMO DATA to see how the tab works.
Pacing Simulator
What happens if you change your behavior? Adjust the sliders.
Inputs
0%50%100%
1h84h168h
1h8h16h
12040
HaikuMixedAll Opus
Quick Presets
TIGHT
Stretching is possible but margins are thin.
Per Hour
1.7%
/hr available
Per Session
1.7%
avg if even
Daily Cap
12.6%
if linear
Effective Burn ×
2.4×
vs all-Haiku
PROJECTED TRAJECTORY
NOW EXHAUSTION @ 38h RESET
Recommendation: Limit sessions to Sonnet-only. Reserve any Opus use for the final session before reset.
Saved Scenarios
No saved scenarios yet. Adjust sliders, then click SAVE PRESET.
Profile Compare
Side-by-side usage profiles · select up to 3
Select Profiles to Compare
Attribute Comparison
Normalized across 6 dimensions
PROFILE MATCH
Select profiles above to see personalized recommendations.
Compare up to 3 profiles · click any profile card to load settings into Simulator
Codex Delegation
Phase 2 primary coding tool · route here when Claude budget tightens
⚙ ROUTING STRATEGY
→
Codex is your Phase 2 primary coding tool (Google AI Pro). When Claude weekly burn exceeds pace, route scaffold · refactor · lint · boilerplate to Codex.
Jules = fallback only. Save Claude for orchestration, planning, and final polish passes.
Codex · Tasks This Week
0/wk
▬ no data yet
Claude Tokens Saved · est.
0%
vs running all tasks in Claude
Hand-off Streak
0days
consecutive days routing to Codex
Routing Matrix
Task type → Codex vs Claude vs Jules
Logged Tasks — 0 entries
No tasks logged yet. Click + LOG TASK to add one.
Path Decision Tally
→ Codex0
→ Claude0
→ Jules0
Codex share —
Broker Router
Intent budget allocation · Claude vs Codex cross-service advisor
Incoming Intents
CLAUDE