Travis CI stalled. Semaphore is where teams landed.
After the pricing changes, cut concurrency, and slowed development, teams moved their CI elsewhere. Semaphore is the fast, actively developed, secure alternative β open source, fairly priced, and built for how teams ship today.
Why teams left Travis CI
These are the reasons developers cite across Hacker News, Reddit, and engineering blogs β and how Semaphore answers each one.
The pricing change that broke open source
The 2020 plan ended reliable free open-source builds; maintainers called the cost “astronomical for non-enterprise usage” and moved projects elsewhere.
Forced .org β .com migration, cut concurrency
The move gave little notice and didn’t carry settings over; some maintainers went from 5 free concurrent jobs to 2 on paid β an immediate performance cut.
The 2021 secrets leak
Reported as a flaw that “exposed secrets for thousands of open source projects.” As one dev put it, secrets are one thing a CI system “has to do perfect.”
Stalled after acquisition & layoffs
Following senior-engineer layoffs, many users felt the product stopped evolving β and quietly migrated to actively developed alternatives.
Slow builds and limited concurrency
Queue waits and capped parallel jobs pushed teams toward platforms with more headroom and unlimited parallelism.
Missing modern CI features
No dashboards, flaky-test detection, monorepo support, promotions, or artifact store; Windows only in limited beta; no SLAs or SOC 2.
Why teams choose Semaphore over Travis CI
Fast, actively developed, and fairly priced β the modern home for CI that Travis users switch to.
Fast builds, no limits
Autoscaling parallelism and optimized pipelines replace Travis’s queues and capped concurrency β so feedback is fast even under load.
Autoscaling parallelism Β· no job-queue delaysActively developed & secure
Open source, images refreshed weekly, and new AI-native capabilities shipping regularly β with the security controls Travis lacked.
SOC 2 Type II Β· ISO 27001 Β· scoped secretsFair, predictable pricing
$15 of free usage every month, then pay only for what you run β no per-seat fees, no idle costs, no credits to ration.
$15 free/mo Β· pay-as-you-go Β· no per-seatTell your agent what you want. It runs your CI.
Semaphore is the CI built for coding agents. Through its MCP server and Claude Code + Codex plugins, your agent sets up CI, inspects runs, debugs failures, and re-triggers jobs β natively. Travis was never built for this era.
- β MCP server: agents read logs & test output programmatically
- β Agent maintains and optimizes your pipeline as code grows
- β Fully open source β read it, run it, self-host it
- β Scoped secrets & job isolation for safe agent access
Your .travis.yml converts in minutes.
Travis pipelines are YAML β so migration is quick. Let your coding agent read your .travis.yml and rebuild it as a clean, versioned Semaphore pipeline.
Estimate your build-time savings
Faster machines and autoscaling parallelism shorten feedback loops. Drag to match your team β this is an illustrative estimate, so measure your own repo for exact numbers.
Illustrative only β the speed-up depends on your workload, caching, and parallelism. The most reliable comparison is running your own pipeline on both.
The differences that actually matter
The five things teams tell us tip the decision. Full feature matrix is available as a download below.
| What you care about | Semaphore | Travis CI |
|---|---|---|
| Pricing | $15 free/mo, pay-as-you-go | Pricing change ended free OSS builds; limited credits |
| Speed & concurrency | Autoscaling parallelism, no queue | Slow builds, limited concurrency |
| Active development | Open source, shipping weekly | Stalled after acquisition & layoffs |
| Security | SOC 2 Type II, scoped secrets, isolation | 2021 secrets-exposure incident |
| Features | Dashboards, flaky detection, monorepo, promotions | Thin feature set; Windows in beta |