When agents start writing real code, the hard part moves downstream. This week’s devtools funding skews toward the plumbing teams hit once AI is in production: observability infrastructure that can actually handle agent traces, and evaluation setups that tell you whether an agent is behaving or just confidently wrong. The launches mirror that too, with new execution layers and frameworks for teams juggling multiple agents in live systems. Here are the funding wins and launches we saw in the devtools startup world over the last 7 days.
🏆 DevTools Funding Wins
Baz Bags $9M To Govern AI Code At Planning Time
Guy Eisenkot and Nimrod Kor founded Baz after scaling Palo Alto Networks’ Cloud AppSec business, and they are now applying that playbook to AI‑generated code. The San Francisco company has closed a $9M seed extension co‑led by Battery Ventures and boldstart ventures, taking total funding to $17M. Baz plans to use the capital to extend its suite of coding agents that sit inside engineering workflows to detect, patch, and validate bugs and vulnerabilities before code hits production, with its new Planner product intervening at the planning stage. 🎉 Congrats to the team, read the full announcement here.
Runpod Raises $100M and Hits $1B Valuation
Zhen Lu built Runpod to be a single destination where AI developers can experiment, train, fine‑tune, run inference, and scale multi‑node workloads without stitching together multiple tools. Lu announced last week that they have raised $100M in growth investment led by Summit Partners at a $1B valuation, and says more than one million developers now use the platform. The fresh capital will fund continued investment in the full‑lifecycle AI development platform and expand engineering and developer relations teams to serve the next wave of AI builders. 🚀 Huge milestone for Runpod, read more on Runpod’s blog.
8090 Labs Lands $135M to Push Enterprise AI Coding Forward
Chamath Palihapitiya founded 8090 Labs in January 2024 to give corporate programming teams an AI coding agent that produces production‑quality software with enterprise controls such as audit trails. The company has closed a $135M Series A led by Salesforce Ventures with participation from WndrCo, Craft Ventures, and angels including Palo Alto Networks CEO Nikesh Arora and Quora CEO Adam D’Angelo. With the raise, Palihapitiya is stepping in as CEO to lead development of the Software Factory product full time. 🧠 This is enormous for the company – find out more from Chamath here.
Arato Secures $10M To Test AI Apps Before They Go Live
Shahar Erez and Hilik Paz previously built and sold Stoke Talent to Fiverr, and they are now tackling a similar validation problem for AI applications through arato.ai. The two‑year‑old Tel Aviv startup has closed a $10M seed round led by TLV Partners with participation from Jibe Ventures and several former CTOs and executives. Arato’s platform simulates how users interact with AI applications so developers can surface failures, risks, and compliance issues before deployment, using text, image, voice, and business data to run thousands of scenarios. 🎯 Well done on the raise Arato! Find out more about Arato here.
🚀 New DevTools Launched
StarSling Ships Self‑Driving CI Runners For GitHub Actions
The team at StarSling (YC X25) has made its AI‑native Runners generally available after running over one million CI jobs. The runners are a drop‑in replacement for ubuntu‑latest that claim to make GitHub Actions up to 6x faster and 13x cheaper, with AI agents that continuously scan workflows, logs, and telemetry and open optimisation PRs to fix caching, parallelise steps, and restructure workflows. Teams like Better Auth and Mastra report significantly faster builds after switching. ⚡️ Nice work from Yonas Beshawred and Daniel Worku! Find out more about StarSling Runners here.
Modelence Brings Mobile Apps To Its Chat‑To‑App Builder
Aram Shatakhtsyan and Eduard Piliposyan, PhD launched Modelence’s App Builder to generate full‑stack web apps from a prompt, and they have now added native mobile app support. Users can describe what they want and get a working iOS or Android app that shares the same auth, backend, and codebase as the web version, aimed at domain experts and technical operators who need production‑grade internal tools on phones. The output is real code on an open‑source framework that can be inspected, extended, and taken elsewhere. 📱 Congrats to the Modelence team on the Product Hunt launch. See the launch page.
Humalike Introduces Behavioral APIs For Humanlike AI Agents
Marti Carmona Serrat and Mateusz Jacniacki met through Y Combinator’s co‑founder matching platform and set out to fix the social awkwardness of AI agents in group chats. They launched Humalike with seven behavioral APIs for turn‑taking, theory of mind, norms, persona, social memory, social signals, and social observability to give agents social skills for group interactions. The platform is model‑ and stack‑agnostic, with one‑shot integrations for Hermes, WhatsApp, and Telegram. 👏 Nice! Check out the launch and support Humalike on Product Hunt.
Feyn Open‑Sources Pulpie, A Cheap Web‑Cleaning Model
Shreyash Nigam and Bhavnick Singh Minhas built Feyn to train custom models on customers’ data that improve over time, and they have now open‑sourced Pulpie for cleaning web pages. The model extracts main content from raw HTML while stripping ads, navigation, sidebars, and other boilerplate, which the team says makes up 70% of a typical page and matters when pages become pre‑training data or retrieval context. Pulpie claims to deliver high‑quality extraction at about one‑twentieth the cost of leading model‑based extractors. 🧹 Good to see this go open source. Read the writeup.
📆 Upcoming Events
DevRelCon NYC 2026 | July 22–23, 2026 | Industry City, Brooklyn, NYC | Talks and workshops on AI developer tooling, agentic platforms, DevEx, and developer adoption strategies. 🎤
WeAreDevelopers World Congress 2026 | 8–10 July 2026 | Berlin, Germany | 500+ speakers across 20+ stages covering AI-era engineering, agentic systems, platform engineering, security, and developer tooling. 15,000+ builders expected. 🌍
WeAreDevelopers World Congress North America 2026 | September 23–25, 2026 | San José, CA | 500+ speakers across 20+ stages covering AI-era engineering, agentic systems, platform engineering, security, and developer tooling. 🌉
Got launch news or funding to share? Send it to Natalie Harper.
Want more details on the opportunities here or want to register your resume with us, reach out to Becca Combe MIRP CertRP and Louise Ogilvy directly.
Keywords: agents, evaluation, infrastructure, funding, launches, coding, security, production
