Varion CI Autopilot · Live comparison

Challenge Varion against your CI.

Run your own CI savings analysis, then inspect direct measured competitor evidence. Varion shows the wins, the smaller advantages, and the cases where a competitor wins.

✓ Live CI analysis✓ Direct BuildBuddy + RunRight evidence✓ Required CI coverage protected✓ No cherry-picked universal claim

DIRECTLY EXECUTED COMPETITORS

Direct benchmark evidence

Measured September 3, 2026. Scenario-specific results, not a universal 50% claim.

DIRECT MEASUREDBuildBuddy

61.86% lower runner time

Three identical required CI checks: BuildBuddy 55.71s → Varion 21.25s.

BuildBuddy ran all three checks; Varion retained one verified primary plus two status-preserving proxies.

DIRECT MEASUREDBuildBuddy

38.94% lower runner time

Same-VPS self-hosted cache test: BuildBuddy 27.43s → Varion 16.75s.

Both warm runs had 82/82 cache hits. BuildBuddy used its official self-hosted image.

DIRECT MEASUREDBuildBuddy

BuildBuddy wins this case

Single warm Cloud RBE/cache job: BuildBuddy 18.17s vs Varion 19.23s.

Shown intentionally so the comparison page does not hide competitor wins.

DIRECT MEASUREDRunRight v1.6.2

0% cheaper runner found

Detected current cost $96.55/mo; cheapest GCP recommendation $193.03/mo.

RunRight measured 88.2% p95 CPU and ~7.92 GiB p95 memory, so keeping the current runner was the rational result.

What we can claim: Varion directly measured 61.86% lower runner time than BuildBuddy in the three redundant-check benchmark. What we do not claim: that Varion always saves 50% more on every workload.

YOUR PIPELINE

Now challenge Varion with your CI.

Paste or upload the configuration below. The public tester analyzes configuration as data and shows current cost, Varion savings, competitor-method context and the exact candidate rewrite.

INPUT

Your pipeline

Upload CI file

CONNECT

Load directly from your CI platform

Credentials are used for this connection only and are not stored by the public tester.

Platform Repository / project Access token

Paste CI configuration

Filename Monthly CI runs Avg job minutes € / runner minute

Verified history / runner evidence optional

Advanced mode: paste Varion shadow-run evidence to count runner arbitrage, superseded-run and impact-routing savings. Unverified evidence is never counted as safe savings.

Test intelligence optional

Upload standard JUnit XML reports from recent CI runs. Varion detects flaky tests from measured history and balances tests across parallel workers using measured timings.

JUnit XML reports Parallel workers

Configuration is analyzed as data. This public tester does not execute uploaded repository code or store repository credentials.

WHY IT SAVES

Optimization evidence

Every action includes the detected evidence and whether verification is still required.

CANDIDATE CHANGE

Review the exact rewrite.

Generated changes remain candidates until shadow verification passes.

No automatic rewrite was generated for this pipeline.

COMPARISON

Same input. Clearly labeled evidence.

Method	Evidence type	Projected cost	Projected saving

LIVE_MEASUREDActual comparable execution was measured for this exact result.

METHOD_REPRODUCTIONPublic optimization method/capability reproduced; not the proprietary service itself.

REFERENCE_ONLYInformational reference; never treated as a measured comparison.

A broad “60% more savings” superiority claim remains locked until a sufficiently broad set of direct measured competitor runs meets the verification gate with equivalent required outputs.