CoralCloud Get Started

Team

Coral Team Members

Engineers, researchers, and builders shaping the future of AI agents.

Portrait of Caelum Forder

Caelum Forder

Founder | CTO

Caelum Forder is the Technical Founder and CTO of Coral AI Labs. After building AI infrastructure at IBM Watson, working on AI safety at Conjecture, and serving as a founding engineer at CAMEL-AI, he designed CoralOS to scale intelligence through agent coordination. His team has delivered leading results across coding, general-purpose agent systems, and scientific autoresearch, and commercialized the work through Coral Code.

Portrait of Peter Carroll

Peter Carroll

Founder | CEO

Peter Carroll is the Founder-Operator and CEO of Coral AI Labs. After bootstrapping a game studio to 30 staff and millions in revenue, he funded Coral’s early development and assembled the team around Caelum Forder’s engineering vision. He leads company strategy, ecosystem growth, and commercialization while contributing to research on multi-agent systems.

Portrait of Suman Deb

Suman Deb

AI Solutions Architect

Suman Deb is a Founding AI Solutions Architect at CoralOS, helping enterprises move multi-agent systems into production. With eight years in AI and an MSc from the University of Aberdeen, he leads client deployments, including a multi-agent platform for asset management, and built Coraliser to turn MCP servers into Coral-compatible agents.

Portrait of Emmett Childs

Emmett Childs

Technical Staff

Emmett Childs is a Member of Technical Staff in Applied AI, turning Coral research and infrastructure into production-grade agentic products. Built Refinery, a Codex-first Coral application that helps ChatGPT and Codex users review agent memory, sessions, skills and files through coordinated specialist agents. Background in data infrastructure and institutional digital asset products.

Questions

FAQs

Something else on your mind? Email hello@coralprotocol.ai

Is it secure? How do you ensure security?

Yes. Our platform is designed with enterprise-grade security in mind. Just one example is where our agents run in isolated, scoped sessions to ensure complete data separation between users. This architecture eliminates the risk of cross-user data exposure, unlike other solutions where data can leak.

How scalable is the platform?

Most multi-agent systems rely on a central planner, which quickly becomes a bottleneck. Coral uses a graph-based architecture, enabling agents to communicate directly and work in parallel without routing everything through one controller.

This design dramatically improves throughput, scalability, and cost-efficiency. Research on similar semi-centralized systems (e.g., Anemoi, Ren et al., 2025) shows that direct agent-to-agent collaboration reduces bottlenecks, improves plan adaptation, and boosts performance — even outperforming stronger baselines on benchmarks like GAIA.

With Coral, agents handle smaller, specialized tasks and coordinate in real time, giving you:

  • High concurrency powered by Coral’s Kotlin-based runtime
  • Better reasoning quality through continuous collaborative refinement
  • Safer, more predictable scaling by adding agents, not bigger models

See benchmark results: https://gaia.coralprotocol.org/

What level of expertise is required to use it?

You don’t need to be an AI expert. The platform is designed to be accessible for beginners while still offering advanced customization options for technical teams. Most users can build and deploy agents with minimal training.

Can we use our own agents?

Absolutely. You can bring your own agents and connect them easily into Coral. Get started today here: https://docs.coralprotocol.org/welcome

How fast can we get onboarded?

Depending on your complexity and requirements, full integration may take a couple of days, but basic setup takes most teams a few hours.

What ongoing support will we need?

You won’t need to contact us for every update. You can add, edit, or retire agents directly through the platform. We provide optional support for anything you need, but day-to-day changes are self-service and easy to manage.