Software teams keep solving the same delivery problem in different scripts. Ozone puts reusable Tekton tasks, visual workflows and release controls in one place - so the next application can borrow the plumbing.
Writing code keeps getting easier. CloudBees has built its business around the harder question: what should be allowed to reach a customer?
A bruising security test gave Simon Bennetts a reason to build a free web scanner. Keeping it free turned out to be a different kind of engineering problem.
A passing test can be a splendid alibi. SeaLights maps software changes to the tests that actually touch them, giving teams a sharper answer to the question: is this release ready?
Software teams can automate thousands of checks and still distrust the result. mabl sells a way to make testing easier - and its own engineers offer a useful lesson in knowing when to stop.
Brave cut builds from two hours to 15 minutes; Snowflake cut the upkeep. EngFlow’s wager is that developer productivity depends on the machinery between writing code and shipping it.
PostHog had 16 setup steps and 14 caveats. Flox helped replace them with a shared environment - and is betting that the same discipline can travel from a developer’s laptop to an AI agent and into production.
A copied API key can outlive the job it was meant to do. Defakto replaces that lingering permission with short-lived workload identity, bringing an old security problem into the age of AI agents.
A NetSuite consultant had a weekend to move a year of work into production. Salto’s bet is that the settings businesses live by deserve the same review, history and release discipline as software code.
A developer's nuisance grew into a company with a large ambition: make the credentials behind modern software easy to find, change, and govern before one forgotten key causes a very public mess.
The Cambridge software company built its business on the awkward work of moving Salesforce changes into production. Now it wants an AI teammate to build those changes, too - with the same checks on the way out.
The Australian developer-tools company won demanding engineering teams with a split-the-stack model: Buildkite coordinates the work, while customers choose where it runs. Now it is trying to turn that architectural compromise into a broader software-delivery platform.
The old bargain was simple: write more tests or accept more bugs. Meticulous is betting that recorded behavior, deterministic browsers and a startling amount of parallel compute can break that bargain - just as AI agents make verification the slowest part of shipping software.
Upsun hides the plumbing beneath modern applications, then gives developers a full-stack rehearsal room for every change. Its wager is that the next cloud winner will sell fewer dashboards and more confidence.
Arga Labs builds real-world sandboxes for AI agents and agent-facing software. Every time a team opens a pull request, Arga spins up a temporary staging environment that mirrors production, redeploying only the services that changed and routing everything else to prod. It runs code and agents against resettable API 'twins' of third-party services like Stripe, Slack, GitHub, and Gmail, so teams can surface edge cases, simulate failures, and test at scale without touching live data. A YC P26 company founded by Phillip Li and Akira Tong, Arga reached roughly $40K in monthly recurring revenue within seven weeks of launch.
Attune is a San Francisco developer-tools startup (YC X25) building drop-in build acceleration for engineering teams. Its tool, invoked by prefixing existing commands with 'hurry' (e.g. 'hurry cargo build'), delivers up to 20x faster builds through incremental compilation, remote execution, and distributed caching - with zero migration and no new config language to learn. Attune started with Rust and is expanding to Docker, npm, GitHub Actions, Swift, and Python. It also maintains an open-source project for fast, secure Linux package publishing. Founded by Xin Ding (CEO) and Eliza, who worked together for three years at FOSSA on build systems.
AutoRABIT is an enterprise DevSecOps and release-management platform built for the Salesforce ecosystem. It combines CI/CD automation (ARM), static code and security analysis (CodeScan), data and metadata backup and recovery (Vault), and security posture management (Guard) so teams in regulated industries - banking, financial services, healthcare, and government - can ship Salesforce changes quickly while keeping security and compliance intact. Founded in 2015 and bootstrapped for its first five years, the company is headquartered in San Francisco and serves customers including ABN AMRO, Cigna, and CyberArk.
Zelar (ZelarSoft) is a cloud-native engineering and consulting firm that helps banks, telcos, oil & gas, and government teams adopt Kubernetes, DevOps, and AI without the usual pain. Founded in 2018 and led by CEO Vasu Maganti, the company pairs hands-on services - cloud migration, SRE, security, and data engineering - with its own platforms: Klusternetes for self-service Kubernetes multi-tenancy, OpenOps for production-ready GKE stacks, and Cokpit for agentic AI DevOps. A Google Cloud Premier Partner with offices across the US, Canada, India, and the UAE, Zelar bets that most teams want cloud-native outcomes, not cloud-native homework.
GitLab is an AI-powered DevSecOps platform that brings the entire software development lifecycle - planning, source code management, CI/CD, security, and deployment - into a single application. Born as an open-source alternative to GitHub in 2011 and run as one of the world's largest all-remote companies, GitLab now serves enterprises and millions of developers, trades on Nasdaq under GTLB, and is pushing into agentic AI development with GitLab Duo.
JFrog is the company behind the universal software supply chain platform anchored by Artifactory, the binary repository that lets engineering teams store, secure, and ship every artifact - packages, containers, and now AI models - from a single source of truth. Founded in Israel in 2008 and headquartered in Sunnyvale, California, JFrog (NASDAQ: FROG) serves thousands of enterprises and crossed $531.8 million in revenue in 2025 while pushing its 'Liquid Software' vision of continuous, secure software flow.
Netlify is the platform that taught the modern web to ship fast. Founded in 2014 by Mathias Biilmann and Christian Bach, it pioneered the Jamstack architecture and a Git-driven workflow that turned every push into a production deploy. Today, more than five million developers and companies like Twilio, Unilever, and Peloton run on Netlify's global edge network - and the company is reinventing itself again around AI agents that ship code on behalf of humans.
Rainforest QA is an AI-powered, no-code software testing platform that helps product teams run regression and functional tests without building or maintaining brittle test frameworks. Founded in 2012 in San Francisco and backed by Y Combinator, the company has executed more than 42 million tests for over 10,000 startups and product teams.
Semgrep is a San Francisco application-security company that builds a unified AppSec platform (Code/SAST, Supply Chain/SCA, and Secrets) used by engineering teams at Dropbox, Figma, Snowflake and others. Born out of an open-source tool originally written at Facebook, it lets security teams write code-like rules and ship them through CI - cutting false positives and pushing fixes back to developers.

Zelos Cloud is a San Francisco-based B2B data platform for mission-critical systems, founded in 2023 by three former Tesla and Neuralink engineers. The platform unifies data collection, remote command-and-control, real-time visualization, and automated testing for firmware and industrial systems - solving the tooling chaos that plagues teams building hardware that cannot afford to fail. Backed by Y Combinator (S23) and Human Capital, Zelos Cloud is building the operating layer that firmware and embedded systems engineers have always needed but never had.

Autosana AI is the first agentic QA platform for iOS, Android, and Web. Founded by Yuvan Sundrani and Jason Steinberg in June 2025 and backed by Y Combinator (S25 batch), Autosana replaces brittle test scripts with AI agents that navigate apps like real users, described in plain English. The company raised $3.2M in February 2026 from YC, Pioneer Fund, Phosphor Capital, DeVC, 468 Capital, and angels including Paul Graham. Revenue is growing 100%+ month over month, and the platform now protects apps serving over 100 million daily active users. The name comes from the Latin 'autosana' — to heal oneself — reflecting tests that evolve alongside the product rather than breaking on every UI change.