FoundryLabs, Inc.
E2B provides open-source, secure cloud infrastructure purpose-built for AI agents, giving every enterprise AI agent its own cloud computer with real-world tools to perform multi-step tasks safely, transparently, and flexibly.
At a Glance
- Fortune 100 and Fortune 500 enterprises
- AI researchers and research institutions
- AI model developers
- Individual software developers and founders building AI agent applications
- +4 more
AI Tools by FoundryLabs, Inc.
(1)
E2B
Cloud Sandboxes for AI Agents
Discussions
No discussions yet
Be the first to start a discussion about FoundryLabs, Inc.
Latest News
E2B Raises $21M Series A to Give Fortune 100 Cloud for AI Agents
88% of Fortune 100 Companies Sign Up for E2B Platform
Czech Startup E2B Raises $11.5M Seed Round to Build Cloud Infrastructure for AI Agents
How Manus Uses E2B to Provide Agents With Virtual Computers
Products & Services
Cloud-based isolated virtual machine environments for executing AI-generated code securely. Powered by Firecracker microVMs with full isolation, starting in less than 200ms with no cold starts. SDKs available for Python and JavaScript/TypeScript.
Sandbox version supporting graphical interfaces and VNC, providing cloud desktops with GUIs for graphical application operation.
Integration with the Model Context Protocol for LLM integration.
Open-source alternative to proprietary systems like Claude Artifacts for building and hosting applications.
Market Position
E2B positions itself as The AI Agents Cloud and the foundation for Cloud 2.0, differentiating from general-purpose cloud computing by providing specialized cloud runtime specifically for AI agents. Competes with general AI infrastructure providers (AWS Bedrock, Google Vertex AI, Azure), specialized code execution platforms (Riza, Daytona, Modal), and agent framework ecosystems. Key differentiators include: fastest sandbox startup times (<200ms), open-source approach, enterprise-grade security with Firecracker microVMs, LLM-agnostic design, and focus on the infrastructure layer rather than full agent hosting. Positioned as essential middleware between AI models and production environments, with 88% Fortune 100 adoption validating product-market fit.
Leadership
Founders
Vasek Mlejnsky (Václav Mlejnský)
Co-Founder and CEO with 10 years of software engineering experience focused on developer tooling and AI-native infrastructure. Originally from Czech Republic, studied at Charles University Faculty of Mathematics and Physics. Previously worked on iPhone game development, computer vision projects, serverless function tools, search engines for developers, and no-code backend builders. Moved to San Francisco in 2021 via Pioneer accelerator. Has collaborated with co-founder since 6th grade.
Tomas Valenta (Tomáš Valenta)
Co-Founder and CTO. Founder of DevBook (predecessor to E2B). Studied at Gymnazium Uherske Hradiste. Background includes iPhone game development, computer vision, and various developer tools. Originally from Czech Republic, childhood friend and long-time collaborator of Vasek Mlejnsky.
Executive Team
Vasek Mlejnsky (Václav Mlejnský)
Co-Founder and CEO
10 years of software engineering experience focused on developer tooling and AI-native infrastructure. Previously built iPhone games, computer vision projects, and various developer tools.
Tomas Valenta (Tomáš Valenta)
Co-Founder and CTO
Founder of DevBook. Computer science background with experience in game development and developer infrastructure.
Board of Directors
Founding Story
Founded in 2023 by childhood friends Vasek Mlejnsky and Tomas Valenta who had been collaborating since 6th grade. Originally started as DevBook, focused on creating interactive documentation and playgrounds for developers. In March 2023, following the release of GPT-3.5, they faced a critical problem while building their coding agent - spinning up VMs took too long and existing infrastructure wasn't suitable for AI agents. They pivoted from DevBook to E2B (English to Bits), shifting focus from hosting full agents to providing the infrastructure layer - secure cloud sandboxes for code execution. The company moved headquarters from Prague to San Francisco to be closer to the AI ecosystem, though recently reopened an office in Prague.
Business Model
Revenue Model
Usage-based API pricing with subscription tiers. Revenue generated from per-second sandbox runtime charges based on CPU and RAM allocation, plus tiered subscription plans.
Pricing Tiers
One-time $100 usage credit, community support, up to 1-hour sandbox session length, up to 20 concurrent sandboxes, 10 GiB free storage. No credit card required.
Everything in Hobby tier, customizable Sandbox CPU and RAM, up to 24-hour sandbox session length, up to 100 concurrent sandboxes, 20 GiB free storage. Payments via Stripe.
Custom solutions with BYOC (Bring Your Own Cloud), on-premise deployment, role-based access control, workload isolation, data residency compliance, special pricing.
Target Markets
- Fortune 100 and Fortune 500 enterprises
- AI researchers and research institutions
- AI model developers
- Individual software developers and founders building AI agent applications
- Data science teams
- Enterprise technology companies
- Deep research agents conducting research on large datasets
- AI data analysis and visualization with secure data exploration
- Coding agents executing generated code with I/O and terminal commands
- Vibe coding: providing code runtime for AI-generated applications
- Reinforcement learning: running and evaluating reward functions
- Computer use: providing secure virtual cloud computers for LLMs
- Hugging Face
- Perplexity
- Manus
- Groq