Zencoder announced the public beta of Zentester, an AI-powered agent that transforms end-to-end (E2E) testing from a bottleneck into an accelerator. This breakthrough enables development teams to move from “vibe coding” to production-ready software by accelerating quality assurance and providing instant verification within developer workflows. While AI coding assistants have revolutionized code generation, the gap between writing code and shipping reliable software remains vast. Teams need faster feedback loops, and E2E testing – the final verification that software actually works – continues to be a manual, brittle process that can add days or weeks to release cycles. Zentester sees and interacts with applications as users do—clicking buttons, filling forms, navigating flows, and validating both UI state and backend responses. The agent can take scenarios in plain English without wrestling with scripting frameworks. This brings comprehensive E2E testing directly to the engineer’s fingertips—both in their IDE through Zencoder’s existing integrations and in CI/CD pipelines via Zen Agents for CI. It enables five mutually supportive use cases: Developer-Led Quality, QA Acceleration, Quality Improvement for AI Coding Agents, Healing Tests, and Autonomous Verification.
Mosaic Agent Bricks platform automates agent optimization and tuning without the need for labeled data
Many enterprise AI agent development efforts never make it to production and it’s not because the technology isn’t ready. The problem, according to Databricks, is that companies are still relying on manual evaluations with a process that’s slow, inconsistent and difficult to scale. Databricks launched Mosaic Agent Bricks as a solution to that challenge. The Mosaic Agent Bricks platform automates agent optimization using a series of research-backed innovations. Among the key innovations is the integration of TAO (Test-time Adaptive Optimization), which provides a novel approach to AI tuning without the need for labeled data. Mosaic Agent Bricks also generates domain-specific synthetic data, creates task-aware benchmarks and optimizes quality-to-cost balance without manual intervention. Agent Bricks automates the entire optimization pipeline. The platform takes a high-level task description and enterprise data. It handles the rest automatically. The platform offers four agent configurations: Information Extraction; Knowledge Assistant; Custom LLM; Multi-Agent Supervisor. Databricks also announced the general availability of its Lakeflow data engineering platform. Lakeflow solves the data preparation challenge. It unifies three critical data engineering journeys that previously required separate tools. Ingestion handles getting both structured and unstructured data into Databricks. Transformation provides efficient data cleaning, reshaping and preparation.
Vanta’s GRC AI agent uses program context to proactively detect inconsistencies between policy-defined service level agreements and the outcomes of continuous testing, flag mismatches and suggest fixes
Cybersecurity compliance startup Vanta has launched Vanta AI Agent, a new agent designed to handle end-to-end workflows autonomously across a company’s entire compliance program. The new agent contextually guides organizations through key tasks, accurately identifies issues and inconsistencies humans might miss and proactively takes action on their behalf, while keeping governance, risk and compliance teams informed and in control. The agent uses program context to offer timely support and surface issues before they become costly errors. It reduces human error by taking on manual, time-consuming tasks and, in doing so, frees teams to focus on higher-value work that builds trust while strengthening their security and compliance posture. In addition to its core functions, Vanta AI Agent also generates clear and actionable policy change summaries, streamlining the process of updating compliance documentation during annual reviews. The result reduces the need for manual input, allowing teams to focus on strategic decision-making. Other features of the agent include the ability to proactively detect inconsistencies between policy-defined service level agreements and the outcomes of continuous testing. When mismatches occur, the agent flags them and suggests fixes, helping teams address issues before they escalate into audit risks. Vanta AI Agent further simplifies information retrieval by answering policy and compliance-related questions in real time. Teams can quickly access critical details such as password requirements, vendor risk management information and compliance with standards like Service Organization Control 2 and the Health Insurance Portability and Accountability Act.
Zencoder’s UI testing AI agent imitates how humans behave when interacting with web applications by combining images with snapshots and generating test artifacts to capture the expected visual and functional outcomes
Zencoder has announced a public beta for Zentester, its new end-to-end UI testing AI agent. Zentester imitates how humans behave when interacting with web applications, such as navigating the layout, and identifying and using interactive elements. It does this by combining images (screenshots) with DOM (snapshot) information. As it runs through test scenarios, it generates test artifacts that capture the actions performed and the expected visual and functional outcomes. According to the company, these tests are designed to be maintainable over time and less prone to having issues when an application changes. It also automatically follows end-to-end testing best practices, such as proper wait strategies, error handling, and test isolation.
Deepgram’s Voice Agent API combines speech-to-text, text-to-speech, and LLM orchestration with contextualized conversational logic into a unified architecture to enable deploying real-time, intelligent voice agents at scale
Deepgram has announced the general availability of its Voice Agent API, a single, unified voice-to-voice interface that gives developers full control to build context-aware voice agents that power natural, responsive conversations. Combining speech-to-text, text-to-speech, and LLM orchestration with contextualized conversational logic into a unified architecture, the Voice Agent API gives developers the choice of using Deepgram’s fully integrated stack or bringing their own LLM and TTS models. It delivers the simplicity developers love and the controllability enterprises need to deploy real-time, intelligent voice agents at scale. Deepgram’s Voice Agent API provides a unified API that simplifies development without sacrificing control. Developers can build faster with less complexity, while enterprises retain full control over orchestration, deployment, and model behavior, without compromising on performance or reliability. Deepgram’s Voice Agent API also provides a single, unified API that integrates speech-to-text, LLM reasoning, and text-to-speech with built-in support for real-time conversational dynamics. Capabilities such as barge-in handling and turn-taking prediction are model-driven and managed natively within the platform. This eliminates the need to stitch together multiple vendors or maintain custom orchestration, enabling faster prototyping, reduced complexity, and more time focused on building high-quality experiences. The platform enables model-level optimization at every layer of the interaction loop. This allows for precise tuning of latency, barge-in handling, turn-taking, and domain-specific behavior in ways not possible with disconnected components.
Groq’s custom Language Processing Unit (LPU) architecture, designed specifically for AI inference enables it to handle memory-intensive operations like large context windows at lower cost compared to general-purpose GPUs
Groq became an official inference provider on Hugging Face’s platform, potentially exposing its technology to millions of developers worldwide. The Hugging Face integration extends the Groq ecosystem providing developers choice and further reduces barriers to entry in adopting Groq’s fast and efficient AI inference. Groq’s assertion about context windows — the amount of text an AI model can process at once — strikes at a core limitation that has plagued practical AI applications. Most inference providers struggle to maintain speed and cost-effectiveness when handling large context windows, which are essential for tasks like analyzing entire documents or maintaining long conversations. Independent benchmarking firm Artificial Analysis measured Groq’s Qwen3 32B deployment running at approximately 535 tokens per second, a speed that would allow real-time processing of lengthy documents or complex reasoning tasks. The company is pricing the service at $0.29 per million input tokens and $0.59 per million output tokens — rates that undercut many established providers. Groq offers a fully integrated stack, delivering inference compute that is built for scale, which means we are able to continue to improve inference costs while also ensuring performance that developers need to build real AI solutions. The technical advantage stems from Groq’s custom Language Processing Unit (LPU) architecture, designed specifically for AI inference rather than the general-purpose graphics processing units (GPUs) that most competitors rely on. This specialized hardware approach allows Groq to handle memory-intensive operations like large context windows more efficiently. By becoming an official inference provider, Groq gains access to the vast developer ecosystem of HuggingFace with streamlined billing and unified access. Amazon’s Bedrock service leverages AWS’s massive global cloud infrastructure, while Google’s Vertex AI benefits from the search giant’s worldwide data center network. Microsoft’s Azure OpenAI service has similarly deep infrastructure backing. However Groq says, “As an industry, we’re just starting to see the beginning of the real demand for inference compute. Even if Groq were to deploy double the planned amount of infrastructure this year, there still wouldn’t be enough capacity to meet the demand today.”
Peymo’s hybrid banking platform uses AI agents that seamlessly integrate fiat, crypto wallets, tokenised assets, and embedded finance into a single stack and identifies the smartest route, timing, and format for each transaction
Peymo Ltd, a UK-based FinTech, has launched the world’s first AI-powered multi-hybrid bank — a digital finance platform that seamlessly integrates fiat banking, crypto wallets, tokenised assets, and embedded finance into one unified system. Built on proprietary modular architecture, the platform enables users to manage GBP, EUR, crypto assets, and branded debit cards in one place, while enterprises can integrate full banking functions via simple APIs. “We make complex finance invisible,” said Tomas Bartos, Founder of Peymo. “By fusing AI, fiat, crypto and embedded finance into a single stack, we’re delivering the next generation of banking — and it’s ready today.” Branded as “Peymo AI – Smarter Banking for Every User,” the platform delivers powerful AI through a voice-first interface with continuous listening, instant intent recognition, and multimodal confirmations to enable secure, hands-free banking. Behind the interface, five specialised AI agents monitor user behavior, track market activity, optimise payments, and ensure asset protection in real time. A built-in smart referral engine identifies potential users within a network, dispatches personalised invitations via WhatsApp or voice, and tracks referral success. Operationally, Peymo’s AI powers instant onboarding in under five seconds, continuous transaction monitoring for faster clearances and real-time deposits, and self-improving system code through usage-based AI feedback to keep the platform fast, compliant, and lean. As a true hybrid financial engine, Peymo’s AI also helps users navigate their entire portfolio — from crypto to fiat, gold to tokenised assets — identifying the smartest route, timing, and format for each transaction, ensuring efficiency, transparency, and full control. Its autonomous architecture supports scalable growth by identifying high-value B2B leads and ideal embedded finance partners who can adopt Peymo’s wallets, KYC tools, cards, or payment systems at scale. Simultaneously, human-sounding voice agents engage users directly — offering guidance, upsell suggestions, and personalised support to unlock unused features, deliver premium upgrades, and execute activation nudges, with the system scalable to millions of tailored interactions per hour.
Algolia’s MCP Server enables LLMs and agentic AI systems to interact with APIs, retrieve, reason with, and act on real-time business context at scale through a standards-based, secure runtime
Algolia announced the release of its MCP Server, the first component in a broader strategy to support the next generation of AI agents. This new offering enables large language models (LLMs) and autonomous agents to retrieve, reason with, and act on real-time business context from Algolia, safely and at scale. Bharat Guruprakash, Chief Product Officer at Algolia. “By exposing Algolia’s APIs to agents, we’re enabling systems that adapt in real time, honor business rules, and reduce the time between problem and resolution.” With this launch, Algolia enables an agentic AI ecosystem where software powered by language models is no longer limited to answering questions, but can autonomously take actions, make decisions, and interact with APIs. The MCP Server is the first proof point in a long-term roadmap aimed at positioning Algolia as both the retrieval layer for agents and a trusted foundation for agent-oriented applications. With the Algolia MCP Server, agents can now access Algolia’s search, analytics, recommendations, and index configuration APIs through a standards-based, secure runtime. This turns Algolia into a real-time context surface for agents embedded in commerce, service, and productivity experiences. Additionally, Algolia’s explainability framework with its AI comes along for the ride for enhanced transparency. More broadly, agents can: Retrieve business; Make Updates Freely; Chain decisions across workflows. With the MCP Server and upcoming tools, Algolia is eliminating friction in the development of agentic AI systems—empowering developers to increasingly: Define agent behaviors around Algolia’s APIs; Rely on Algolia’s safety scaffolding; Compose agents that span systems.
Typedef turns AI prototypes into scalable, production-ready workloads by managing all the complex properties of mixed AI workloads through a clean, composable interface using APIs, relational models and serverless tech
Typedef Inc., turning AI prototypes into scalable, production-ready workloads that generate immediate business value, has come out of stealth mode with $5.5 million in seed funding. With a new purpose-built AI data infrastructure for modern workloads, Typedef is helping AI and data teams overcome the well-documented epidemic affecting the bulk of enterprise AI projects – failure to scale. The solution is built from the ground up with features to build, deploy, and scale production-ready AI workflows – deterministic workloads on top of non-deterministic LLMs. Typedef makes it easy to run scalable LLM-powered pipelines for semantic analysis with minimal operational overhead. The developer-friendly solution manages all the complex properties of mixed AI workloads like token limits, context windows, and chunking through a clean, composable interface with the APIs and relational models engineers recognize. Typedef allows for rapid, iterative prompt and pipeline experimentation to quickly determine production-ready workloads that will demonstrate value – then realize that potential at scale. Typedef is completely serverless bypassing any infrastructure provisioning or configuration. Users simply download the open-source client library, connect their data sources and start building their AI or agentic pipelines with just a few lines of code. No complex setup, no infrastructure to provision, no brittle custom integrations to troubleshoot.
Rackspace Technology and Sema4.ai launch industry’s first scalable enterprise AI agent solution
Rackspace Technology announced a strategic alliance with enterprise AI agent innovator Sema4.ai (“Sema4”). This collaboration integrates the Foundry for AI by Rackspace (FAIR™) services and Rackspace’s application management expertise with Sema4.ai’s advanced ‘SAFE’ AI Agent Platform, combining the strengths of both companies in artificial intelligence, cloud and systems integration to accelerate the adoption of secure, enterprise-grade AI solutions. The partnership will enable the rapid deployment of scalable, production-ready AI agents across enterprise functions with robust governance, transparency, and security protocols as their core foundational elements. Through this new collaboration, businesses can design and deploy custom AI agents tailored to specific use cases, with seamless integrations across key functions such as HR, finance, customer support, sales, and operations. Customers will also gain access to full AI lifecycle management, including built-in observability and a centralized control plane, enabling scalable, secure, and efficient deployment. These agents can: Operate independently or collaboratively, offering shared functionalities such as natural language understanding, workflow automation, and advanced document processing. Help streamline and enhance the automation of high-value tasks. Automate existing business processes, which interpret standard operating procedures (SOPs) and runbooks, eliminating the need for complex prompt engineering.