Discover. Fix. Prevent.
The complete CI reliability platform.
These are the capabilities of the CI Brain — the parts that know (memory), decide (triage, scoring), and act (suppression, blocks, burn-in, bisect). The chat assistant and auto-insights tie them together. Release Confidence Score, CI waste attribution, failure clustering, agent-ready context — one platform.
The full CI reliability lifecycle — one platform.
Discover
Stop chasing phantom failures
A test that passes 9 times out of 10 wastes more engineering time than one that consistently fails. Drape tracks pass/fail patterns and flags flaky tests before they erode trust in CI.
- Flakiness scoring based on historical pass/fail rates
- Branch-aware detection — distinguish environment flakes from code bugs
- Group flaky tests by detection commit for batch triage
- Track CI time wasted on retries caused by flakes
Discover
One number that tells you: can I ship?
Release Confidence aggregates first-attempt success rate, test health, code coverage, and flakiness into a single 0–100 score. Drop below your threshold and you know before cutting the release — not after.
- Weighted scoring from 4 inputs — tuned to what actually predicts release quality
- Automatic penalties for high flake rates or declining coverage
- Trend tracking — see your score improve over time as you fix root causes
- Per-repository scores — compare health across your entire org
Discover
Every wasted CI minute, categorized and ranked
Drape breaks down CI waste into five categories — flaky retries, real failures, infrastructure issues, retry overhead, and queue wait time — then ranks the worst offender jobs so you know exactly what to fix first.
- Five waste categories — see where your CI budget actually goes
- Top offender jobs ranked by savings potential
- Trend charts — track whether your CI efficiency is improving or regressing
- Retry overhead and queue wait broken out separately — not hidden in averages
Discover
Stop fixing symptoms. Fix root causes.
When 15 tests fail, it's usually 1–2 root causes. Drape clusters related failures by log similarity and ranks them by impact — so your team fixes the underlying problem, not each symptom individually.
- N-gram frequency clustering — groups failures with similar error signatures
- Impact score ranking — occurrences × affected jobs
- NEW badge for first-time failure patterns
- Cross-job correlation — find infrastructure issues that span multiple pipelines
Prevent
Coverage that only goes forward
Percentage targets are a blunt instrument. Teams game the number, and a 0.1% drop is noise. Drape enforces coverage at the line level: if a PR drops coverage on changed files, it gets flagged before it merges.
- Line-level tracking, not just percentages
- Commit-to-commit comparison — see exactly what changed and when
- Detect regressions on changed files, not the whole codebase
- Supports Cobertura, LCOV, and Go coverage — no proprietary format required
- Cross-repo visibility for monorepos and multi-service architectures
Fix
Suppress flaky tests and jobs without losing track
Suppression isn't an excuse to ignore problems. Drape enforces suppression hygiene: expiration dates, linked fix tickets, and automatic staleness alerts ensure flaky tests and jobs get fixed, not forgotten.
- Suppress flaky tests and flaky CI jobs — same workflow, same accountability
- Branch-pattern matching — suppress on
mainonly,feature/*, or everywhere - Expiration dates — suppressions auto-deactivate if not renewed
- Ticket URL linking — every suppression points to a fix ticket
- Staleness tracking — surface items suppressed for 30+ days with no progress
Discover
Diagnose failures across your entire test suite in seconds
Google built TestGrid. Meta built Probabilistic Flakiness Scores. Spotify built Odeneye. Dropbox built Athena. Every engineering org at scale ends up building an internal test health dashboard. Drape gives you one out of the box.
- Grid view — tests as rows, commits as columns, status at every intersection
- Spot patterns instantly: scattered red means flaky, solid columns mean infrastructure
- Duration heatmaps — see which tests are getting slower over time
- Filter by branch, status, or test name — drill into what matters
- Retry detection — multiple runs on the same commit grouped automatically
| Mar 12 | Mar 11 | |||||||
|---|---|---|---|---|---|---|---|---|
| 15:00 | 13:00 | 10:00 | 08:00 | 00:00 | 20:00 | 16:00 | 12:00 | |
8c7d8c7d | 1b4a1b4a | e6f3e6f3 | c2d5c2d5 | f9a1f9a1 | d4e8d4e8 | b7c2b7c2 | a3f1a3f1 | |
| test_api_timeout | ||||||||
| test_checkout_flow | ||||||||
| test_db_reconnect | ||||||||
| test_export_csv | ||||||||
| test_search_results | ||||||||
| test_user_login | ||||||||
Everything else your CI pipeline needs
Fix
Agent-Ready Context
Structured API and MCP integration give AI agents failure context, flakiness scores, and coverage gaps — so they fix issues instead of retrying blindly.
Prevent
Vulnerability Lifecycle & Security Debt
Ingest SARIF from Grype or Trivy. Track time-to-remediation, SLA compliance by severity, and security debt trends. Two-scan resolution protection prevents premature closure.
Prevent
Code Analysis
SARIF ingestion from ESLint, Semgrep, golangci-lint, and more. Baseline your state, block new violations, ratchet toward zero.
Discover
Activity Feed
Automatic regression and resolution detection across flaky tests, coverage, security, lint, and duration. Fingerprint-based deduplication.
Discover
AI Failure Classification
Automatically triage every failure — real bug, flaky test, or infrastructure issue. Classify failures as new regressions or previously known issues so developers know instantly whether a red build needs attention.
Prevent
PR Insights
Rich PR comments with new flakes, coverage regressions, CVEs, lint violations, and duration spikes — updated on every push.
Discover
Merge Cycle Time & DORA Bands
Track PR open-to-merge time with p50/p90 percentiles. Automatic DORA band classification — Elite, High, Medium, Low — so you speak the same language as your VP of Engineering.
Fix
Burn-in & Bisect Campaigns
Suspect a test is flaky? Burn it in — run it 50 times on the same commit. Found a regression? Bisect it — Drape walks the commit history with binary search to find the breaking change. Trigger from the UI, API, or automatically.
Discover
Team Ownership (CODEOWNERS)
Drape ingests your CODEOWNERS file and syncs GitHub teams. See which team owns the most flaky tests, the most security debt, or the lowest coverage — and route accountability to the right people.
The infrastructure every scaling team builds
Google, Meta, Spotify, and Dropbox each invested years building internal CI reliability infrastructure. Your team should be shipping features, not rebuilding their tooling.
Google — TestGrid
Open-source dashboard for visualizing test results across thousands of builds. Maintained by a dedicated team.
Meta — Probabilistic Flakiness
Bayesian scoring system monitoring millions of tests. Built to create accountability and drive reliability improvements.
Spotify — Odeneye
Grid visualization that reduced flaky test rates from 6% to 4% — just by making the data visible to teams.
Dropbox — Athena
Automated build health management system. Detects, suppresses, and tracks flaky tests across all repositories.
Drape gives you what these teams spent years building — deployed in 5 minutes.