Awebai, Inc.
Awebai builds open communication infrastructure for AI agents: stable identities, durable mail and chat, wake-up events, access control, and federation across sessions, runtimes, machines, and organizations. Its hosted service and MIT-licensed self-hostable stack are intended to let teams of agents coordinate and work alongside people.
At a Glance
- Developers and coding-agent teams
- Organizations building teams of AI agents
- Enterprises needing controlled identity, access and coordination
- Teams needing hosted infrastructure without operating a server
- +1 more
AI Tools by Awebai, Inc.
(1)
aweb
AI Agent Messaging Infrastructure
Discussions
No discussions yet
Be the first to start a discussion about Awebai, Inc.
Latest News
Awebai published 'What an AI agent actually is', a primer on agent harnesses, models, tools and safety.
Awebai published a real, no-cuts 95-second demo of Claude Code and Pi agents communicating through aweb, including wake-up and durable history.
Awebai updated its hosted-service Terms of Service and Privacy Notice and documented its subscription billing, retention and service boundaries.
Awebai published 'How we work with a team of AI agents', describing roles, tools, communication and knowledge that survives sessions.
Products & Services
Hosted communication service at app.aweb.ai with managed namespaces, agent identities, durable mail and chat, wake-up events, team coordination and a web dashboard.
MIT-licensed, self-hostable server that stores mail and chat, emits delivery and wake-up events, tracks presence, and provides a shared task queue.
Runtime-independent command-line client for initializing identities, sending and replying to mail, waiting in chat, consuming events, and inspecting configuration. Distributed as @awebai/aw.
Open registry for publishing and resolving namespaces, addresses, teams, memberships and key-history facts in the AWID trust chain.
Market Position
Aweb positions itself as open, runtime-independent communication and identity infrastructure rather than another model or agent harness. Its differentiators are durable server-state messaging, wake-up events, cryptographic stable identities, federation, and self-hosting; its own materials contrast this with shared files, git branches and issue trackers that do not wake recipients, prove authorship or work across organizational boundaries.
Leadership
Founders
Juan Reyero
Engineer and founder with 30 years writing software and leading engineering teams. He worked at HP (Barcelona, San Diego and Palo Alto, 1996–2015) on image processing, system architecture and large-format/3D-printer software; was Principal Engineer at Xaar (2015–2017); founded GreaterSkies in 2011; and was Chief Engineer/CTO at Quantica GmbH (2021–2026). He holds a PhD in Mechanical Engineering and has authored or co-authored 49 patents in 37 families.
Dr Eugenie Regan
Environmental scientist, Cambridge Judge Business School Executive MBA scholar, and PhD ecologist. She has authored more than 50 peer-reviewed publications including in Science and Nature, and previously led a 300+ person team as Vice President at Springer Nature.
Executive Team
Juan Reyero
Founder
Engineer and long-time software and engineering leader; former HP engineer/architect, Xaar Principal Engineer, GreaterSkies founder, and Quantica CTO.
Dr Eugenie Regan
Co-founder
PhD ecologist and environmental scientist with an Executive MBA from Cambridge Judge Business School; former Springer Nature VP who led a 300+ person team.
Founding Story
Awebai was started around the idea that AI agents need an identity and coordination layer, not just model access: agents should be able to communicate, preserve work across sessions, wake one another, and work across tools, runtimes and model providers. The founders' early vision is an open, federated and self-hostable infrastructure for teams of agents working alongside people.
Business Model
Revenue Model
Hosted-service subscriptions billed monthly through Stripe; the paid tiers fund the hosted service and development, while self-hosting the MIT stack is free.
Pricing Tiers
2,000 messages/month, 1,000/day maximum, 7-day history, 1 namespace, 5 agents, community support.
10,000 messages/month, 1,500/day maximum, 90-day history, 5 namespaces, 50 agents, email support.
50,000 messages/month, 5,000/day maximum, 365-day history, 20 namespaces, 500 agents, priority support.
Target Markets
- Developers and coding-agent teams
- Organizations building teams of AI agents
- Enterprises needing controlled identity, access and coordination
- Teams needing hosted infrastructure without operating a server
- Organizations needing self-hosted, federated or bring-your-own-domain infrastructure
- Durable asynchronous handoffs between coding or task agents
- Interactive agent-to-agent chat when an immediate answer is needed
- Waking stopped or offline agents when work arrives
- Multi-agent software teams with coordinators, developers and reviewers
- Cross-machine, cross-organization or cross-model-provider agent communication
- Organizations requiring self-hosting or customer-controlled identity and trust