This week’s developer tools are taking us from vibes to visibility. From smarter code reviews to battle-tested AI agents, the focus is clear: better insight, less guesswork. We’ve got launches, funding news, and one platform bringing real observability to LLMs.
Got something we should feature next week? Hit reply or send it our way.
🏆 Funding Wins
Bito AI Banner (Source: LinkedIn)
Bito’s AI Agent Levels Up the Code Review Game with $5.7m Seed Extension
Bito has raised a $5.7M seed extension to tackle one of the biggest bottlenecks in software development: code review. While most AI tools focus on writing code, Bito’s agentic AI reviews, reasons, and fixes it; helping teams ship safer, faster! Already reviewing over 10,000 PRs a week for companies like Gainsight and PubMatic, the platform claims to cut PR merge times by 89% and reduce regressions by 34%. The round was led by Vela Partners with backing from NextView, Eniac, and others.
📣 Shoutout to co-founders Amar Goel, Anand Das and Mukesh Agarwal. Read more about their mission here.
Diversion Secures $12M to Build the GitHub for Game Devs
Version control meets game development in Diversion, the Y Combinator-backed startup building collaboration tools for studios working with massive 3D assets. Now armed with $12M in total funding (including a fresh $4.5M Seed+ round led by Konvoy) the team plans to expand support for AAA studios and onboard even more creative teams. With 15,000 devs already using the platform and 20% monthly growth, Diversion is fast becoming the go-to for game versioning!
🎮 Congrats to CEO Sasha Medvedovsky and the team. Find out more here and book a demo.
Assisterr Raises €2.4M to Put AI Creation in Everyone’s Hands
London-based assisterr is building the infrastructure for a decentralised AI creator economy, and just secured €2.4M to keep going! Their no-code platform lets anyone build, deploy, and monetise small language model agents without relying on centralised tools or APIs. The round includes backing from Google for Startups, Outlier Ventures, Web3.com and others. With 4.7M users and 24,000+ agents already live, they’re scaling fast, focusing on use cases from DeFi analytics to tokenised art tools.
“From the start, our mission has been to provide AI for all, not just for Big Tech. The current AI landscape is incredibly centralised, expensive, and, frankly, out of reach for most individuals and smaller businesses. We’re changing that by making AI creation as easy as building a website, giving power, ownership, and monetisation back to the creators.” – Nick Havryliak (CEO and Co-Founder)
🚀 Big congrats to co-founders Nick Havryliak and Dmytro Dimenko. Check them out here!
🚀 3, 2, 1… Launches
Build, Test, Deploy — All in One Browser with Operative.sh
Operative.sh is now in public beta, offering a browser-based AI agent that not only writes code but tests it too. As it builds, it launches a real browser, scrolls pages, clicks elements, checks console errors, and captures screenshots, so your app works straight out of the gate. As the founders put it:
“We’ve always loved tools like Cursor, Cline, and Loveable — they make it super easy to generate code. But what we found was that using them still left you in the loop: prompt feature request, get output, test it, click fix, paste errors or describe problem, and repeat.”
🎯 Try it now at operative.sh, watch the launch video here and congrats to Christopher Settles and Erik Q. on the launch!
Janus Banner (Source: LinkedIn)
Janus Brings Crash-Test Simulation to Conversational AI
Janus (YC X25) has launched to help AI teams stress-test their agents before users ever touch them. Founded by Shivum P. and Jet Wu, the platform simulates thousands of multi-turn conversations in minutes. It flags hallucinations, tool-call failures, biased responses, and other risky outputs. Instead of waiting for issues in production, Janus surfaces them upfront with performance benchmarks and root cause insights.
📍Congrats to the Janus team on the launch! Building with agents? Book a 15-min demo
dbt Labs Debuts Fusion Engine for Faster Data Development
dbt Labs has overhauled its data transformation platform with Fusion, its fastest engine yet, built in Rust for massive speed gains. It powers all core dbt tools and cuts parse times by up to 30x. Also new: a VS Code extension, dbt MCP server for AI integrations, and dbt Canvas, a visual modelling workspace. CEO Tristan Handy called it:
“the biggest launch in the history of the company.”
🔧 Read the full announcement and explore what’s new here.
Emergent Wants to Be Your AI-Powered CTO
Emergent is live and gaining momentum as an agentic “vibe-coding” platform that turns plain-language prompts into full-stack, production-ready apps. Founded by twin brothers Mukund Jha and Madhav Jha, Emergent handles planning, coding, testing, and deployment in a single flow. Over 10,000 apps have already been built on the platform, from Slack bots to full SaaS tools.
✨ Explore what’s possible on Emergent and watch their intro video. Congrats to the team!
Retool Agents Bring Autonomy to Business Workflows
Retool has launched Retool Agents, a toolkit for building autonomous AI systems that connect to APIs, reason through decisions, and carry out complex business tasks. They can be triggered by email, chat, or workflows, and monitored with built-in evaluation tools. It’s part of Retool’s broader push into AI-powered internal tools, with flexible pricing and free monthly usage to get started.
💼 Find out more from Keanan Koppenhaver (Technical Product Marketing Manager) on Product Hunt (plus give it an Upvote whilst you’re there!). Hats off to the team for this big release!
Simplex Banner (Source: LinkedIn)
Simplex Builds Web Agents for Legacy Systems
Simplex is out of stealth and helping AI teams interact with legacy portals without relying on APIs. It lets developers automate workflows like form submissions or invoice edits directly through constrained web agents. Built for reliability and transparency, it is already being used in production by teams that need precise control over agent behaviour.
📎 Check it out at simplex.sh and congrats to co-founders Shreya Karpoor and Marco Nocito.
BitPatrol Finds Secrets in Code Where Others Miss
BitPatrol (YC X25) scans source code for secret leaks such as tokens and API keys, using machine learning to understand developer intent. It was founded by Christoper Lambert, a former Stripe engineer and top-ranked HackerOne researcher. Unlike regex-based tools, BitPatrol cross-references public code to assess exposure and reduce false positives. It’s available via GitHub App for real-time scanning and custom workflows.
🔐 Get started at bitpatrol.io.
HelixDB Blends Graph and Vector Search for RAG and AI Apps
HelixDB is an open-source graph-vector database designed for AI and RAG applications. Created by George Curtis and Xavier C., it supports semantic vectors and graph links in the same system, enabling developers to build faster and more contextual retrieval tools. It also offers type-safe queries and blazing-fast performance.
🧠 Star the project on GitHub or visit helix-db.com to dig in. Congrats to the team on the public release!
🔦 DevTool of the Week: Traceloop
Building with LLMs today can often feel like trial and error. Outputs shift without warning, prompts break silently, and debugging turns into a mess of spreadsheets and Slack threads. Traceloop was born out of this frustration.
Founded by ex-Google and Fiverr engineers Nir Gazit and Gal Kleinman, Traceloop gives developers visibility into how AI agents behave in the wild. Their open-source SDK, OpenLLMetry, traces prompts, responses, failures, and quality metrics across frameworks like MCP, Agent2Agent, and AGNTcy. It’s already trusted by IBM, Cisco, Dynatrace, and Miro.
Last week, Traceloop launched publicly AND announced $6.1M in funding to expand its platform.
“Prompt engineering shouldn’t be a guessing game or have to rely on ‘vibes’ to be successful…It should be like the rest of engineering – observable, testable, and reliable. When we bring the same rigor to AI that we expect from the rest of our stack, we unlock its full potential.” (Nir Gazit)
With over 50,000 weekly SDK installs and half a million monthly downloads, Traceloop is fast becoming essential for teams deploying production-scale AI.
🎯 Star them on OpenLLMetry and Hub on GitHub and give them a follow on LinkedIn and Twitter. Explore more at traceloop.com
—————————————————————————————————-
Thanks for reading. We love shining a light on the builders making developer tools better, faster, and more reliable. If you spotted a new tool, launch, or raise worth shouting about, we want to hear about it. Message one of the team below.
Louise Ogilvy (Recruitment Director) – UK/USA
Becca Combe MIRP CertRP (Lead Recruiter) – USA
Ben Chipchase (Lead Recruiter) – UK
Natalie Harper (Administrator)
Looking for your next DevTools role? See the latest Jobs.
