🍊 faf drift

Context-drift makes you pay.
.faf keeps you drift-free. 🏁

AI without context makes bad decisions. Those decisions compound.
Detect and measure the real cost of context-drift in your projects.

How to Run Analysis

Project Analysis

Current .faf Score
29% 🟡
Drift Events Detected
4
Estimated Time Lost
8-10 weeks
Future Risk Level
HIGH

Drift Cascade Timeline

How context-drift compounds over time

🔴
Framework Week 0
Added React integrations
Cost: 2-4 weeks
↓
🔴
Database Week 8
Migration to new ORM
Cost: 1-2 weeks
↓
🟡
Styling Week 12
CSS Modules → Tailwind
Cost: 1 week
↓
🔴
Framework Week 18
Vite integration added
Cost: 2-4 weeks

The Cascade Effect

Each drift event makes the next one more likely. Without project context, AI suggests solutions that don't match your architecture, leading to more drift.

4 events across 18 weeks = 8-10 weeks lost

Real Project Comparison

Same team. Different context strategy.

❌ WITHOUT .FAF FROM DAY 1

faf-dev-tools
29% 🟡
Drift Events: 4
Time Lost: 8-10 weeks
Risk Level: HIGH

What Happened:

  • 🔴 Auth migration (JWT → Firebase → Supabase)
  • 🔴 State management changed twice
  • 🟡 Styling framework replaced
  • 🔴 Framework integrations added reactively

Each change made without project context. AI couldn't warn about conflicts because it didn't know the architecture.

☑️ WITH .FAF FROM DAY 1

claude-faf-mcp
99% 🏆
Drift Events: 0
Time Lost: 0 weeks
Risk Level: LOW

What Happened:

  • ✅ Consistent architecture from day 1
  • ✅ AI understood project context always
  • ✅ Decisions aligned with existing stack
  • ✅ Zero drift detected across 131 commits

.faf provided project DNA from the start. AI made context-aware suggestions that matched the architecture perfectly.

The Difference

Projects WITH .faf from day 1: ZERO DRIFT
Projects WITHOUT .faf: 8-10 weeks lost to context-drift

Calculate Your Own Drift Cost

Don't like our numbers? Adjust them.
A lost day is too many.

Your Estimates

Lower them. Higher them. The truth won't change.

2.5 weeks
4 days
4 days
3 weeks
1.5 weeks
3 fixes/event

Your Project Cost (4 Drift Events)

Event 1: Framework 3 weeks
Event 2: Database 1.5 weeks
Event 3: Styling 0.8 weeks
Event 4: Framework 3 weeks
Subtotal (Base Events) 8.3 weeks
Cascade Effect +2.4 weeks
TOTAL TIME LOST 10.7 weeks

Reality Check

53 working days
lost to context-drift
$64,200
cost at $150/hour developer rate

10.7 weeks = 10 sprints lost. How do you explain this to leadership?

With .faf From Day 1

0 weeks
No drift. No rework. No excuses.

Projects with .faf from the start show ZERO architectural drift.
Prevention is free. Rework costs 10.7 weeks.

The Truth Won't Change

Move those sliders. Make yourself feel better.

Every drift event costs time. Cascade effects are real.
Even half our numbers prove it.

A lost day is too many.
Stop hiding. Start preventing. Get .faf.

How We Calculate Time Lost

Industry-standard estimates, validated by real projects

📊 Base Estimates (Per Event)

  • Auth Migration: 2-3 weeks JWT → Firebase → Supabase requires code rewrites, testing, security audits
  • State Management: 3-5 days Redux → Zustand needs store refactoring, component updates
  • Styling Framework: 3-5 days CSS Modules → Tailwind = every component restyled
  • Framework Integration: 2-4 weeks Adding React/Next/Vite requires build config, migration testing
  • Database ORM: 1-2 weeks Prisma → Drizzle = schema migration, query rewrites, data validation

⚡️ Cascade Multiplier

Each drift event doesn't happen in isolation. When you change auth, you also need to update:

  • State management (sessions)
  • API routes (authentication)
  • Database schema (user tables)
  • Tests (auth mocks)
  • Documentation

Heuristic: 2-5 fix commits per drift event
Example: 4 events × 3 avg fixes = 12 cascade commits

🔍 Detection Method

The CLI analyzes your git history to detect:

  • Package.json changes: Multiple auth/state/styling packages = drift
  • Commit messages: "migrate", "switch", "replace" patterns
  • File patterns: Auth files, store files, style files changed

Calculation: Sum base estimates + cascade effect
Example: Auth (2-3w) + State (3-5d) + Styling (3-5d) = 3-4 weeks → rounded to 8-10 weeks with cascade

✅ Validated By Real Projects

faf-dev-tools
4 drift events detected
Calculated: 8-10 weeks lost
Current .faf score: 29% 🟡
claude-faf-mcp
ZERO drift events
Calculated: 0 weeks lost
Current .faf score: 99% 🏆

Methodology: Conservative industry estimates validated against actual git history analysis. We use the LOWER end of ranges unless cascade effects are detected.

Run Your Own Analysis

Detect context-drift in your projects

1

Install FAF CLI

npm install -g faf-cli
2

Run Drift Analysis

cd /path/to/your/project
faf drift --export drift-report.json
3

Upload Report

Click "Upload Drift Report" above to visualize your results

What the CLI detects:

  • 🔍 Auth drift: JWT → Firebase → Supabase migrations
  • 🔍 State management: Redux → Zustand → Jotai changes
  • 🔍 Styling: CSS Modules → Tailwind → Emotion switches
  • 🔍 Framework: React → Next.js → Vite integrations
  • 🔍 Database: Prisma → Drizzle → TypeORM migrations

Each pattern detected = weeks of rework caused by context-free AI suggestions

Stop Context-Drift Before It Starts

Get .faf from day 1. Give AI your project DNA. Avoid costly drift.

Get .faf Now →