pull_request_#847.diff

@@ -42,7 +42,12 @@ export function ReviewService {

const config = loadConfig();

const client = new ApiClient(config);

- const result = await client.fetch(url);

+ const result = await client.fetch(url, {

+ timeout: 30000,

+ retries: 3,

+ headers: getAuthHeaders()

+ });

return processResult(result);

}

@@ -78,4 +83,15 @@ async function authenticate() {

const token = await getToken();

- return { token };

+ if (!token) {

+ throw new AuthError('No token');

+ }

+ return {

+ token,

+ expiresAt: Date.now() + 3600000

+ };

}

@@ -112,6 +128,9 @@ class PRReviewService {

async splitPR(pullRequest) {

+ // Intelligent grouping by functionality

+ const groups = this.analyzer.group(pullRequest);

+ return groups.filter(g => g.size > 0);

}

@@ -156,2 +175,8 @@ function validateInput() {

const schema = getSchema();

+ if (!schema.validate(input)) {

+ logger.warn('Validation failed', { input });

+ return { valid: false, errors: schema.errors };

+ }

+ return { valid: true };

// AI-powered code review

Review in
minutes,
not hours.

AI splits large PRs into reviewable groups.

Start reviewing ← free to start

// The problem

After 200 lines, reviewers
stop reading.

60%

fewer bugs caught
in large PRs

longer review
cycle time

75%

devs avoid
large PRs

2-3d

avg delay
for reviews

→ AI generates code faster than humans can review. Copilot writes 500 lines in minutes. Your team reviews at the same pace they did a decade ago. The bottleneck isn't generation—it's comprehension.

// How it works

From chaos to
clarity.

before: PR #847 847 lines

1 M auth/login.ts

2 M api/users.ts

3 A components/Header.tsx

4 M hooks/useAuth.ts

... +23 more files

$ review --status

"I'll review this later..."

after: 5 groups ✓ organized

Authentication Flow 84 lines

auth/login.ts, hooks/useAuth.ts

User API Updates 156 lines

api/users.ts, types/user.ts

UI Components 203 lines

components/Header.tsx +4

+2 more groups...

→ Review each in ~10 min

// Features

Intelligent Splitting

Conservative, dependency-aware analysis groups related changes together. Target 25-100 lines per group—the optimal range where reviewers maintain full attention.

AI-Powered Narratives

Each group comes with a story. Understand what changed, why it matters, and what to look for—even in code you've never seen.

Deep Analysis

Security vulnerabilities. Performance issues. Code smells. Testing gaps. AI catches what tired reviewers miss—before production.

Review Tracking

Mark groups as reviewed. Track team progress. Know exactly what's been examined and what still needs attention.

Privacy Mode

Your code, your choice. Enable privacy mode to fetch code on-demand without storing it on our servers. Perfect for sensitive codebases.

Encrypted Storage

When storing code, all diff content is encrypted at rest with unique keys per PR. AI analysis results are also encrypted. Your code stays protected.

// Security & Privacy

Your code is protected.

We built Cleave with security-conscious teams in mind. Choose how your code is handled—from zero-storage privacy mode to encrypted persistence.

privacy_mode.config

Privacy Mode

Code is fetched on-demand from GitHub and never stored on our servers. Only metadata and AI insights are persisted.

✓ Zero code storage

✓ On-demand GitHub fetch

✓ Session-only caching

encryption.config

Encrypted Storage

All code diffs encrypted at rest with unique per-PR keys. AES-256 encryption with secure key wrapping.

✓ AES-256-GCM encryption

✓ Per-PR unique keys

✓ Encrypted AI analysis

data_control.config

You're in Control

Configure data handling per user or per organization. Delete your data anytime. Export everything.

✓ Per-user privacy settings

✓ One-click data deletion

✓ Full data export

SOC 2 Type II (planned)

GDPR Compliant

No code training

// Pricing

Bring your own keys.

You control your AI costs. Use your existing OpenAI, Anthropic, or Groq keys. We charge for the platform, not the tokens.

Free

$0 forever

Get started with basic PR reviews

+ 3 PRs/month
+ 1 org
+ Basic grouping
+ BYOL AI

Get Started

Popular

Individual Contributor

$3.00 /month

Unlimited PRs for solo developers

+ Unlimited PRs
+ 3 orgs
+ AI grouping
+ GitHub sync
+ BYOL AI

Request Access

Team

$5.00 /seat/mo

Collaboration features for development teams

+ Unlimited PRs
+ 5 orgs
+ AI grouping
+ GitHub sync
+ Team workspaces
+ BYOL AI

Request Access

Organization

Custom

Enterprise features with SSO, audit logs, and dedicated support

+ Unlimited PRs
+ Unlimited orgs
+ AI grouping
+ GitHub sync
+ Team workspaces
+ SSO / SAML
+ BYOL AI

Contact Sales

✓ All plans: OpenAI, Anthropic, Groq | Organization: +Bedrock, +Azure, +Vertex

$ cleave --start