June 2026 Summaries

Rogo is an AI platform tailored for financial institutions, enhancing their ability to conduct research, analysis, and complex workflows through the use of sophisticated AI agents. The platform operates on two distinct layers: Claude Managed Agents for orchestrating specialist and generalist tasks, and E2B for executing code in isolated sandboxes to ensure security and auditability. Rogo utilizes Managed Agents for their robust orchestration and post-training capabilities, allowing the company to efficiently manage tasks requiring specialized or general knowledge without compromising performance. The E2B layer provides kernel-level isolation, enabling thousands of concurrent sandboxes to execute code securely and efficiently, which is crucial given the highly regulated nature of Rogo's clientele. The integration of these systems allows Rogo to maintain a high level of security and scalability, essential for their operations with hundreds of financial firms worldwide, while also ensuring rapid startup times and efficient query handling through pre-configured templates and domain-specific network controls.

Partnering with Stripe Projects, E2B aims to simplify the deployment of AI coding agents by eliminating the need for human intervention in infrastructure provisioning and credential management. This collaboration introduces the Agentic Provisioning Protocol, which enables agents to autonomously discover, provision, and authenticate secure sandboxes using Stripe's trust layer, without manual dashboard interactions or API key retrieval. By leveraging Stripe's existing identity and payment verification systems, developers can swiftly provision E2B environments with a single command, synchronizing and rotating credentials seamlessly. Once an E2B sandbox is provisioned, agents can fully utilize the E2B SDK for managing code execution and sandbox lifecycles, aided by prebuilt templates for various coding frameworks. The use of Firecracker microVMs ensures robust isolation, supporting diverse programming languages for extended runtimes. This integration marks a significant step in creating a standardized, auditable, and efficient protocol for automated code execution, with future expansions and collaborations anticipated.

StackAI is a no-code enterprise AI platform founded in 2023 by MIT PhDs Tony Rosinol and Bernard Aceituno, designed to automate document- and process-heavy tasks using agentic workflows that integrate with systems like SharePoint and SAP. As StackAI's customer workflows became more sophisticated, the company required secure, isolated code execution environments that scaled effectively, leading to the integration of E2B's sandbox technology. E2B provides StackAI with isolated virtual machines for running agent-generated code, ensuring compliance in regulated industries like banking, defense, and healthcare. This partnership allows StackAI to focus on developing revenue-generating features by offloading the maintenance of sandbox infrastructure, which is crucial for scaling hundreds of agent workflows in production. E2B supports three main use cases for StackAI: custom code execution in Python or JavaScript, robotic process automation (RPA) for secure desktop interactions, and safe browser operations for web data scraping. StackAI's future plans involve treating sandboxes as core infrastructure, providing each agent with a persistent environment to maintain state across sessions, thereby enabling further growth and innovation.