Semaphore open source 1.6 is almost here. 🚀 It’s been exactly one year since we released 1.5, and a lot has happened since then. Around 494 pull requests later, we’re packaging up a year of new capabilities and sharing them with the community. The big theme this year: AI and automation. The MCP server now ships in both the Community and Enterprise Editions, sem-ai works out of the box, and almost everything in Semaphore is now configurable through the API. Add partial job reruns and RustFS for artifact storage, and it’s a big release. 📖 Read the full breakdown on our blog: semaphore.io/blog/semaphore-… ▶️ Or watch the full product update on YouTube: youtu.be/bX8f_Zd9XBk

Semaphore preflight checks let teams validate branch names before a workflow begins. With sem-ai 0.4.0, you can ask Claude Code to create these checks directly from your development environment. In this example, Claude starts configuring a rule that: • Accepts main and master • Requires feature branches to use the feat/* pattern • Blocks noncompliant branches before the pipeline begins It is a practical way to turn delivery standards into enforceable safeguards while keeping the workflow inside your coding agent. Watch the full sem-ai 0.4.0 demo on the Semaphore blog: semaphore.io/blog/sem-ai-0-4… #CICD #ClaudeCode #DevOps #AICoding #Semaphore

A code map does not need to start with a complex tool. Semaphore engineer Marcos began with structured Markdown files that helped his agents trace paths across microservices. The result was faster task startup and shorter prompts. Watch more: youtu.be/DrYxInyumHM

Code maps give AI agents a clearer view of code paths, services, and dependencies. Semaphore engineers are using them to provide better context, reduce token consumption, and help agents make fewer incorrect assumptions. Full discussion: youtu.be/DrYxInyumHM

Traditional flaky test rules can miss problems that appear across branches or over time. Semaphore is exploring a signal based model where reruns and changing results contribute to a flakiness score. Watch the full update: youtu.be/DrYxInyumHM

AI agents can generate code quickly. But when a test fails, they need complete context to understand what happened and what should change. Semaphore is working on richer test reports that preserve more information from individual test runners, making failures easier for developers and AI agents to interpret. Full product update coming out tomorrow. Stay tuned. #Semaphore #AITesting #CICD #AICoding #SoftwareTesting

The hard part of a Jenkins migration is not just moving pipelines. It is untangling legacy plugins, scripts, dependencies, and versions that may no longer be supported. Try Semaphore: semaphore.io Watch the full talk: youtube.com/watch?v=HzXAdtVr… #Jenkins #CICD #DevOps #DeveloperProductivity

AI agents are only as useful as the context they receive. When a workflow fails, CI/CD can provide the signals agents need to understand the issue, help fix it, and keep developers moving. Watch the full talk: youtube.com/watch?v=HzXAdtVr…

Agentic development shifts the bottleneck from writing code locally to orchestrating agents reliably. CI is becoming the execution layer: a scalable environment to run, test, and govern those workflows. Watch the full talk: youtube.com/watch?v=HzXAdtVr… @BretFisher

Rebuilding a pipeline no longer means rerunning every job in the block. Now Semaphore rebuilds only the jobs that failed. Passing jobs and earlier ones are carried over and reused, keeping your workflow topology intact. Need the full-block rebuild back? There's a new block_rebuild flag in the YAML for that. See the full demo: semaphore.io/blog/rerun-only…

The sem-ai engineers will be joining @BretFisher shortly to talk about AI-native CI/CD. Please join us for the live discussion and questions: youtube.com/watch?v=Vka1TaLr…

Marko Gacesa, Head of Product at Semaphore, shares how our engineers shape up the features at Semaphore CI. Watch the full conversation on CI/CD at Scale with Senors at Scale at: Youtube: youtu.be/nJ4zhHhEiNs?si=Zlzp… Spotify: open.spotify.com/episode/55l…

New homepage. New era of CI. 🚀 Semaphore is now agent-native CI — built for the age of coding agents, not config files. Your agent sets up CI, validates changes before you push, and keeps pipelines fast as your codebase grows. Zero setup. Fully open source. Give our sem-ai a try at semaphore.io

Deployment failed because a CI secret was missing. sem-ai creates the secret, updates the pipeline to use it, validates the config, and reruns the deployment without leaving the terminal. github.com/semaphoreio/sem-a… semaphore.io #CICD #DevOps #Secrets #AI

Don't push just to see if CI passes. sem-ai runs your uncommitted changes in a real CI environment using the same machine and image as your pipeline. Test first, push later. Learn more about sem-ai: github.com/semaphoreio/sem-a… semaphore.io #CICD #DevOps #AI

A flaky integration test failed. sem-ai finds the failed job and reruns only that block. The passing jobs stay untouched. Rerun the failure, not the pipeline. ⚡ Try sem-ai out at: github.com/semaphoreio/sem-a…

"How's our CI doing this week?" Pass rate 71%. Avg run ~1 min. Slow ones: 3.5 min. But the build itself is <30 sec. The rest is queue time. Fix: agent capacity, not the pipeline. CI health check on demand. That's @semaphoreci github.com/semaphoreio/sem-a…

sem-ai can now find your flakiest tests via the API and fix them automatically, at ~$1-1.50 per fix. Flaky test API is now GA. semaphore.io/flaky-test-api-…