The dashboard — today

Five views. One answer: where do I start?

Endure gives you a full picture of your codebase health — from a single health score down to the individual files, their co-change connections, and the patterns your codebase keeps repeating.

endure.codeslick.dev · Overview

Codebase health at a glance

Paste a public GitHub URL or point at a local path. Endure returns a Health Score, severity distribution, and the top debt files — in under 60 seconds.

Health Score 0–100 Severity breakdown Highest debt files Intent coverage

endure.codeslick.dev · Files

Endure Files view — sortable list of all analyzed files with debt score, complexity, churn, and severity

Every file ranked by actual risk

Sort by debt score, complexity, or churn. Filter by severity or directory. Click any file to see its full breakdown and open the co-change graph.

Debt score per file Cyclomatic complexity Churn rate Staleness score

endure.codeslick.dev · Co-change Graph

Endure Co-change Graph — network visualization of files that historically change together, with AI-captured intent tooltip

See the hidden connections

Files that always change together are coupled — whether you know it or not. The graph surfaces these relationships from git history, so you know the blast radius before you touch anything.

Co-change detection Coupling strength AI-captured intent Shared commit history

endure.codeslick.dev · Knowledge Base

Endure Knowledge Base — patterns learned from your codebase, with severity, confidence, and CLI verification commands

Your codebase's own failure patterns

Endure learns recurring risk patterns from your specific git history — not generic rules from a rulebook. Each pattern comes with a confidence score, severity, and a CLI command to verify it manually.

Learned from your history Confidence scoring CLI verification Grows over time

endure.codeslick.dev · Trends

Endure Trends — average debt score over time, showing improvement across analysis runs

Is your codebase getting better or worse?

Every analysis run is stored as a snapshot. Trends shows you whether debt is accumulating or being reduced — correlated to specific dates so you can tie spikes to deploys, refactors, or team changes.

Score over time Per-run history Critical file tracking

The problem

Code is easy to write.
Understanding it months later is the hard part.

Linters catch syntax. Security scanners catch vulnerabilities. Tests catch regressions. Nothing catches the slow erosion of understanding — until a critical incident reveals it.

🌀

Intent is never captured

Every file was written for a reason. Those reasons live in Slack threads, PR descriptions, and the heads of people who left six months ago.

📉

Code drifts from its purpose

Each change made without understanding the original intent is a small bet against the future. Enough small bets compound into architectural collapse.

🔇

Degradation is invisible

Complexity scores tell you what. Churn tells you where. Nothing tells you why something is becoming fragile or what it's supposed to be doing.

⚡

AI makes it exponentially worse

AI-generated code appears without the context that would have existed if a human wrote it slowly. Velocity without understanding is just faster forgetting.

endure · codeslick · research preview

The six pains of codebases without memory

ENDURE

AI Code Maintenance
Intelligence

code that forgets,
breaks.

01 ·

Lost Intent

// no one remembers why

Original purpose vanished with the developer who wrote it.

02 ·

Silent Drift

// last modified: unknown

Code diverges from its assumptions across hundreds of PRs. Nobody notices.

03 ·

Debt Blindness

// unknown risk surface

Can't tell which files are actually risky vs just messy. Guesswork.

04 ·

Knowledge Silos

// bus factor: 1

One person understands this module. They just handed in their notice.

05 ·

Maintenance Roulette

// here be dragons

Every change to high-churn files breaks something in production.

06 ·

Stale Code Anxiety

// do not touch

Untouched for years, still critical. Too risky to change, too costly to ignore.

How it works

Capture once. Detect drift forever.

Endure runs alongside your existing workflow — no new process, no mandatory annotations. It builds understanding progressively from what already exists.

Analyze your repository

Point Endure at any git repository — public GitHub URL or a local path. It scores every file on three dimensions: complexity, churn, and staleness — weighted into a single Debt Score and an overall Health Score for the whole codebase.

Intent is extracted automatically

For high-debt files, Endure asks an LLM: what is this file supposed to do? The answer is self-judged for quality, stored in a database, and attached to future analysis runs. No manual annotation required. 93% average confidence on the first pass.

Co-change connections are mapped

From your git history, Endure finds files that always change together — even when the connection isn't obvious from the code itself. This coupling map is the blast-radius map every engineer needs before touching anything high-risk.

The Knowledge Base learns your failure patterns

Every analysis run feeds into a self-improving Knowledge Base. Endure extracts recurring patterns from your specific codebase history — not generic rules — assigns confidence scores, and surfaces the ones most likely to cause your next incident. Each pattern includes a CLI command to verify it manually right now.

Trends show you if you're winning

Every run is stored as a snapshot. The Trends view shows whether your codebase health is improving or deteriorating — so you can correlate score changes to specific deploys, refactors, or team changes and know if your remediation efforts are actually working.

Built on 6 principles

Not a linter. Not a metrics dashboard.
Something new.

Persistent Intent

Intent is the invariant; code is the implementation. Intent must outlive any single file or function and be versioned as rigorously as code.

Context as Dependency

Missing context causes maintenance failures, not compile failures. Context has a lifecycle: created, consumed, evolved, deprecated.

Inevitable Drift

All systems drift from their original assumptions. Drift is not failure — blindness to drift is. Measurable drift enables corrective action.

Earned Understanding

Understanding cannot be reconstructed from code alone. It must be captured at creation time. It degrades without active preservation.

Antifragility

Systems should improve from maintenance stress. Each fix should reduce future failure probability. Learning must be structural, not just human.

Maintenance as First-Class

Maintenance is not a cost — it's the product learning. Maintenance readiness is measurable. Maintenance velocity is optimizable.

MVP · Research Preview · 2026

Try it on a real codebase.
Tell us what you find.

We built this to validate a hypothesis: that teams desperately need to know where technical debt actually lives, not just that it exists. Use it for free. Share what surprised you.

Open the dashboard — paste a GitHub URL

Public GitHub repositories — no signup, no API key needed.

Or leave your email and we'll be in touch directly — no scripts, no sales decks.

No spam. No sales calls. Just a conversation about your codebase.

TypeScript / JavaScript Python Java Go Any git repository No signup required

Know which files
are about to cause problems.
Before they do.

Five views. One answer: where do I start?

Codebase health at a glance

Every file ranked by actual risk

See the hidden connections

Your codebase's own failure patterns

Is your codebase getting better or worse?

Watch the product work on a real codebase.

Code is easy to write.
Understanding it months later is the hard part.

Intent is never captured

Code drifts from its purpose

Degradation is invisible

AI makes it exponentially worse

Capture once. Detect drift forever.

Analyze your repository

Intent is extracted automatically

Co-change connections are mapped

The Knowledge Base learns your failure patterns

Trends show you if you're winning

Not a linter. Not a metrics dashboard.
Something new.

Persistent Intent

Context as Dependency

Inevitable Drift

Earned Understanding

Antifragility

Maintenance as First-Class

Try it on a real codebase.
Tell us what you find.

Know which files are about to cause problems. Before they do.

Five views. One answer: where do I start?

Codebase health at a glance

Every file ranked by actual risk

See the hidden connections

Your codebase's own failure patterns

Is your codebase getting better or worse?

Watch the product work on a real codebase.

Code is easy to write.Understanding it months later is the hard part.

Intent is never captured

Code drifts from its purpose

Degradation is invisible

AI makes it exponentially worse

Capture once. Detect drift forever.

Analyze your repository

Intent is extracted automatically

Co-change connections are mapped

The Knowledge Base learns your failure patterns

Trends show you if you're winning

Not a linter. Not a metrics dashboard.Something new.

Persistent Intent

Context as Dependency

Inevitable Drift

Earned Understanding

Antifragility

Maintenance as First-Class

Try it on a real codebase.Tell us what you find.

Know which files
are about to cause problems.
Before they do.

Code is easy to write.
Understanding it months later is the hard part.

Not a linter. Not a metrics dashboard.
Something new.

Try it on a real codebase.
Tell us what you find.