LayerNorm
LayerNorm builds Overlay, an open-source workspace and control plane where humans and AI agents share context, knowledge, tools, computers, and workflows. Its stated aim is to amplify human potential rather than replace it, while preserving privacy, openness, control, and simplicity.
At a Glance
- Individual AI power users
- Small and collaborative teams
- Engineering and operations teams using agents
- Organizations needing governed AI workflows
- +2 more
AI Tools by LayerNorm
(1)
Overlay
Open Source AI Agent Workspace
Discussions
No discussions yet
Be the first to start a discussion about LayerNorm
Latest News
Overlay Web shipped ongoing reliability, security, dependency, messaging, and model-catalog updates in the public repository.
Repository updates included React Doctor security fixes, lifecycle email redesign, workspace mentions/DM delivery, and agent-owned PR workflow documentation.
Web tools were routed through ELO/TinyFish and the prior gateway hop was removed, according to the repository change log.
Enterprise runtime and self-hosting packaging was reworked around the self-hosted Convex backend, with EC2 deployment and environment updates.
Products & Services
Open-source web application and server workspace for humans and AI agents, with chat and work modes, persistent memory, knowledge bases, projects, media generation, browser automation, integrations, and scheduled automations. It is available as hosted Overlay, self-hosted software, and enterprise deployment options.
Public-source macOS Apple Silicon desktop client providing a persistent desktop surface, local/cloud data flows, voice transcription, browser and file operations, and permissioned agent execution connected to Overlay Server.
Market Position
Overlay positions itself as an open-source, provider-neutral control plane and shared workspace rather than a single-model chat product: it unifies context, agents, tools, computers, channels, and automations, while offering self-hosting, enterprise controls, and model portability. This differentiates it from closed single-vendor assistants and from narrower point tools for chat, coding, browser automation, or workflow orchestration; its closest competitive set is open agent workspaces, AI employee/automation platforms, and enterprise private AI platforms.
Leadership
Founders
Divyansh (Dev) Lalwani
Biomedical Engineering and Applied Mathematics student/graduate of Johns Hopkins University; built AutoQuill, a macOS voice-dictation application using Flutter, Groq, and Whisper before founding LayerNorm. LinkedIn identifies him as LayerNorm's Founder and CEO.
Executive Team
Divyansh (Dev) Lalwani
Founder and CEO
Biomedical Engineering and Applied Mathematics background at Johns Hopkins University; previously created AutoQuill and now leads LayerNorm and Overlay.
Founding Story
LayerNorm was formally incorporated in February 2026 after Divyansh Lalwani had begun building AI interface software while at Johns Hopkins. The founding thesis, expressed in the Overlay manifesto, is that model providers are becoming interchangeable while the interface and context layer is where users face fragmentation and lock-in; LayerNorm therefore set out to create an open, extensible interface that brings models, tools, files, and workflows together under user and institutional control.
Business Model
Revenue Model
Freemium subscription and usage-budget model: a free hosted workspace, paid monthly plans whose budget is consumed by premium AI and execution features, customizable monthly budgets, one-time/automatic top-ups, and enterprise/private deployment pricing. Self-hosted source is AGPL-licensed, with commercial licensing and managed enterprise deployment options documented.
Pricing Tiers
Unlimited Auto-model messages, notes, chats, files, knowledge, basic AI tools and core flows, and 10 MB file storage.
1 GB storage; premium chat models, Daytona sandboxes, browser tasks, image and video generation, plus Free features.
Custom monthly budget; includes premium models, Daytona sandboxes, browser tasks, image/video generation, and advanced agents/premium workflows.
Governed, private/on-prem or managed-cloud deployment capabilities.
Target Markets
- Individual AI power users
- Small and collaborative teams
- Engineering and operations teams using agents
- Organizations needing governed AI workflows
- Enterprise and regulated customers seeking private or self-hosted deployment
- Developers and platform teams extending an open-source AI workspace
- Human-agent collaboration on research, writing, and project work
- Delegating repeatable workflows and scheduled automations
- Browser-based research and computer-use tasks
- Team knowledge management across files, notes, memories, and projects
- Connecting agents to Slack, Telegram, iMessage, email, calendars, GitHub, Notion, and other work tools
- Private/self-hosted AI workspaces for organizations