The Quiet Power Behind Smarter AI Agents

Not every week in devtools is about big shifts. Sometimes it’s about quiet confidence; smarter infrastructure, better AI agent control, smoother handoffs. The kind of progress that helps teams trust what they’re building. Agent tooling is getting more thoughtful. Not only slick demos, but practical ways to deploy, manage, and shape how agents behave in real systems. The knobs are getting easier to turn. The metrics are getting easier to see. That’s what this week’s developer tools newsletter is about.

Less noise, more clarity.

🏆 Funding Wins

Leapter Locks in €2M To Build a Glass Box for AI Development

Leapter has raised €2 million in pre-seed funding to replace the AI black box with something developers and business teams can actually trust. Led by bm|t with participation from SIVentures and angel investors, the round will help Leapter expand its visual-first platform for AI-driven software design. Rather than generate opaque code, Leapter creates structured diagrams that let cross-functional teams co-design systems, trace decisions, and ensure compliance. Co-founders Oliver Welte and Robert Werner, both ex-VMware and Broadcom, built Leapter to close the trust gap in enterprise AI.

As Werner puts it;

“You can’t just throw AI-generated code over the wall and hope for the best. That’s why we built Leapter around ‘executable blueprints’—to empower the entire team to build systems that are correct by design.”

💎 Learn more about Leapter here. Congratulations Leapter!

Qedma Raises $26M to Tackle Quantum’s Error Problem

Israeli startup QEDMA has raised $26 million in Series A funding to help quantum computing cross the next frontier: error mitigation. The round was led by Glilot+ with IBM, Korea Investment Partners, and TPY Capital joining in. Qedma’s QESEM software detects and suppresses errors in quantum circuits in real time, extending performance on today’s hardware. Co-founders Asif Sinay, Professor Dorit Aharonov, and Professor Netanel Lindner are bridging academic research and industry demands with hardware-agnostic solutions already used via IBM’s Qiskit catalog. With this raise, Qedma plans to grow its 40-person team and push toward real-world quantum advantage.

⚛️ Explore their mission.. Well done Qedma!

TopK Picks Up $5.5M to Rethink Enterprise Search for AI

Founded last year, Prague-based TopK has secured $5.5 million in seed funding to build an AI-native search engine for structured and unstructured data. The round was led by Earlybird and joined by KAYA, Irregular Expressions, and several angels. TopK’s hybrid retrieval platform integrates vector search with traditional filtering, enabling fast and accurate queries across massive datasets. Co-founders Marek Galovic and Jerguš Lejko are on a mission to create the go-to query layer for AI agents and developers alike.

🔎 Read their announcement here! Congrats TopK!

Airweave Snags $6M to Turn App Chaos into Agent Clarity

Airweave (YC X25), has raised $6 million in seed funding to turn everyday app data into usable knowledge for AI agents. The open-source platform connects to tools like GitHub, Notion, and Postgres, automatically structuring data for seamless agent retrieval. Founded by Lennert Jansen and Rauf Akdemir, Airweave supports everything from vector syncing to graph mapping, with minimal code overhead. The new funding will support engineering hires, new integrations, and deeper community engagement across its web UI, SDKs, and REST APIs.

💨 Here more from the team here.. Amazing work Airweave!

🚀 3, 2, 1… Launches

Kusari Launches Inspector to Shift Security Left with AI

Kusari has launched Inspector, a new AI-powered tool that performs real-time security risk analysis whenever a developer opens a pull request. It’s designed to embed security into the development workflow without slowing anyone down. CTO Michael Lieberman says Inspector uses AI and dependency graph analytics to flag vulnerabilities and guide developers through fixes with inline, annotated reports. Developers can also chat with Inspector, clarify findings, and generate a full SBOM automatically. At $10 per seat per month, Inspector aims to bring intelligent, traceable security checks into the dev loop itself.

🥑 Try it out and explore more at kusari.dev and the full introduction to Kusari Inspector here

Portia AI SDK Update Brings Safer Agent Workflows

Fresh off a #1 Product Hunt launch, Portia AI has released a major SDK update focused on transparency, compliance, and developer control. Used to deploy reliable AI agents in regulated industries, Portia’s new features include declarative agent planning, stateful execution tracking, and guardrails that enforce human authorisation. With over 1,000 native tool integrations, the SDK also makes deployment flexible; on your own infra or via Portia’s cloud. Co-founders Mounir Mouawad and Emma Burrows, both ex-Stripe, built Portia to bring structure and oversight to AI agents operating in high-stakes domains.

🤖 Explore the SDK and docs at portialabs.ai and show your support on Product Hunt

Dittofeed Launches Embedded Components for Built-In Messaging

Dittofeed has introduced Embedded Components, a new way for SaaS teams to add messaging features directly into their apps. These include iframes and, soon, headless React components for template editors, workflows, broadcasts, and segmentation. The launch was driven by user demand, after dozens of devs requested embeddable features even before the team advertised them. Co-founders Chandler Craig and Maxwell Gurewitz built Dittofeed as an open-source alternative to Klaviyo and Iterable, now making embedded messaging possible with minimal engineering effort and optional self-hosting. Early adopters are already live, and the team is actively seeking feedback.

💬 Show your support on Product Hunt 📨 Explore the full platform at dittofeed.com

a0.dev Rolls Out WAGMA to Speed Up App Creation

YC-backed a0.dev (YC W25) has launched Phase 1 of its platform, introducing WAGMA: We Are All Gonna Make Apps. Designed for mobile developers using React Native, a0.dev generates ready-to-deploy app projects in minutes. The latest updates include one-click iOS builds, App Store submissions, payments support, and GitHub syncing. Developers can also access agent mode, thinking mode, and Stripe integration. Founded by Seth Setse and Ayomide Omolewa, a0.dev is focused on reducing the time between idea and execution by handling the hard infrastructure work behind the scenes.

📱 Explore Phase 1 and start building at a0.dev

Relari Launches Nuvi to Build Agents with Plain English

Relari has launched Nuvi, a platform that turns natural language specs into testable, production-ready AI agents. Built for Software 3.0, Nuvi helps teams define behaviour in plain English, simulate outcomes, and deploy agents without writing code or configuring prompts. Each agent spec is structured, versioned, and verified to ensure reliable performance. Use cases already include legal automation, mortgage fulfilment, and nonprofit finance coaching. Founded by Yi Zhang and Pasquale Antonante, Relari is also behind open-source tools like Agent Contract and continuous-eval, designed to keep AI systems aligned with user intent.

🧠 Try Nuvi and build your first agent at nuvi.dev

Coder Kicks of Launch Week with Secure, Self-Hosted Coding Agents

Coder has kicked off Launch Week with Coder Tasks, a tool that lets developers deploy coding agents on secure, self-hosted infrastructure. Designed for enterprise teams, Coder Tasks supports both IDE-integrated and headless agents like Claude Code, Cursor, and Amazon Q. Admins can define templates using Terraform to manage permissions, API access, and workspace boundaries. Teams can now work with agents safely, without exposing sensitive files or slowing down workflows. Coder Tasks is available in both Open Source and Premium tiers, starting with version 2.25.

💻 Get started with Coder Tasks and keep in the loop about the rest of the week

🔦 DevTool of the Week: Retell AI

Voice agents have come a long way, but most still struggle with basic things like interruptions or awkward pauses. Retell AI is tackling that head-on with a developer-first platform for building and deploying real-time, human-like voice agents.

Founded in 2023 by Bing Wu, Zexia Zhang, Weijia Y., Evie Wang, and Todd Li, the team came through YC’s Winter ’24 batch and has been shipping fast ever since. Latency clocks in at around 800ms. Developers can bring their own LLMs, spin up agents over WebSocket or SIP, and monitor every interaction with rich post-call analytics.

The team recently announced a partnership with OpenAI to power its agents using GPT‑4o and GPT‑4.1, enabling phone agents that “spin up natural-sounding voice agents to handle real conversations—booking appointments, answering questions, and resolving issues in real time”. As CTO Zexia Zhang said on LinkedIn;

“With every new model release, our platform evolves. OpenAI makes it possible for us to solve harder problems with a simpler stack.”

The results have been amazing. Check out the full article about this partnership here.

Retell also caught our attention last week during their 2nd launch week; it went a little bit like this:

• Day 1 – KB 2.0: Smarter knowledge base recall, up to 50% better accuracy

• Day 2 – Extract Variables in Real Time: Agents now capture and pass details mid-call

• Day 3 – Warm Transfer 2.0: Human detection, whisper, and three-way messages

• Day 4 – Analytics 2.0: Visualise post-call metrics with custom dashboards

• Day 5 – Cartesia Voice: A new lifelike voice option powered by Cartesia

These solid updates really give a clear signal of where Retell AI is heading in terms of giving developers a clear way to build, deploy, and scale voice agents that actually work in the real world.

🗣️ Explore how it works at retellai.com

If this week shows anything, it’s that good developer tools aren’t always loud about it. But they make the work feel steadier. More transparent. Easier to shape without losing control.

Thanks for reading — and if you’re building something in this space, we’d love to hear about it.

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?

Visit devtoolsjobs.com