49 New AI Tools This Week: Agents, Voice, and Safety

Week 4 of building aitoolfinder.ai, and we're seeing some clear trends emerge across the tools being released. More agents, more safety tooling, and a lot of enterprise infrastructure getting real. Let's dig in.

AI Agents Are Becoming Infrastructure

Three tools this week signal where AI agents are headed: Relevance AI is a no-code platform for building multi-step AI workflows that combine language models with data retrieval—think of it as Zapier for AI. Prentis, a new lab from Reid Hoffman and Mark Pincus, is focused on agents that automate routine computer tasks without requiring code. And OpenAI Presence is the enterprise version—AI agents designed to handle voice and chat across business systems.

The pattern here is clear: agents are moving from "cool demos" to infrastructure. We're seeing purpose-built platforms, funded startups, and major vendors all betting on this.

Voice Tech Is Having a Moment

Fish Audio raised $50M to build voice cloning and synthesis for creators and enterprises. Doppely is doing similar work with support for 40+ languages and emotion control. Even Health in ChatGPT now integrates with your medical records for personalized insights.

What's interesting is the split between consumer and professional use cases. Voice cloning used to be a novelty. Now it's infrastructure for content creators, accessibility tools, and healthcare applications.

Enterprise Automation Is Accelerating

Google's Gemini API Managed Agents lets developers build autonomous agents without managing infrastructure—that's a big deal for teams that don't want to run their own backend. Xano is a no-code backend platform with AI automation built in. Devbox creates instant cloud development environments with automatic dependency detection.

Enterprise wants faster deployment and less operational overhead. These tools are solving that problem.

Safety and Compliance Are Growing Categories

Microsoft's new cybersecurity AI model is their first dedicated security model, paired with an agentic system for threat response. Anthropic's Constitutional AI Dashboard lets researchers evaluate and audit safety properties of language models. And there's even a technical case study on AI agent intrusion documenting a simulated breach.

This tracks with regulatory pressure and enterprise demands. If you're building AI systems, safety tooling is no longer optional.

The Open-Source Foundation Is Growing

Helix by DuckDuckGo is an open-source platform for fine-tuning and deploying LLMs. Puter is an open-source OS that runs in your browser with integrated AI. These represent the counter-move to proprietary closed platforms.

Some Niches Worth Watching

Looplift generates travel itineraries with real-time recommendations. ReWords AI edits text inside images while preserving typography. Helix by Factor creates personalized fitness plans based on your genetics. Kawa preps you for job interviews with an AI interviewer.

Vertical AI tools are proliferating. Rather than general-purpose models, we're seeing specialized tools for specific workflows—recruiting, real estate, fitness, design.

One Thing We're Not Adding

A few of the submissions this week were news articles, podcasts, or blog posts rather than actual tools. We don't list those. If it's journalism or commentary about AI, it doesn't belong in a tools directory.

That leaves us with roughly 41 genuine tools this week. Plenty of signal for teams building in this space.

Explore all 49 additions in the full directory, and let us know what you're building.