Deriv Open-Sources Deriv One, Its Self-Hosted Enterprise AI Platform
October 30, 2026
From a skills registry request to a self-hosted AI platform, free for any developer to run.
Dubai, 29th October, 2026 – Deriv has released Deriv One, a self-hosted AI platform, as open source under the MIT licence. It lets anyone at a company use the AI skills their role allows, and any developer runs it. It is free, no vendor account needed.
Access for everyone, with control built in
Deriv One began as CEO Rakshit Choudhary’s request for a skills registry. As agents spread, each team wrote its own skills, and nobody could say who could run which.
“AI shouldn’t be limited to the few people who can set up their own agent. With Deriv One, anyone at the company can use the skills their role allows, and the guardrails are already built in,” said Rotimi Akinyele, newly appointed Chief Technology Officer at Deriv.
AI-first transformation is already deep inside Deriv. Under Choudhary’s leadership, Deriv has set a target of automating 75% of manual workflows by the end of 2026. Deriv One, an in-house layer on top of AI models, is where staff turn their skills into agent tasks. The 221 skills came from 54 people across the company, not one team.
What everyone gets
• Workspace: split panes, outputs as tabs, and cheaper-model routing.
• Skills: reviewed markdown documents, visible by role.
• Agents: their own Slack apps, each with a named owner.
• Lab: describe a tool in plain language, preview and deploy it.
For developers: run it in three commands
“We’re releasing the same workspace, skills catalogue, agent registration and Lab that we run internally. Clone it, point it at your own model gateway, and see how governed AI access works in practice,” Akinyele said.
Point it at any OpenAI-compatible gateway, including local LM Studio or Ollama:
git clone [REPO_URL] && cd platform
npm run setup -- --quick --test-gateway
npm run dev:stack
It includes sandboxing, Docker, Kubernetes manifests and GKE and EKS guides. Deriv’s production cloud plugins, internal integrations and some Lab features are excluded; adopters bring their own gateway, identity provider and cloud.
“Models will keep changing. The hard part is the layer underneath: identity, secrets, skills review and sandboxing. That’s what we’re sharing, so other teams don’t have to rebuild it,” Akinyele said.
Links: GitHub · MIT licence · Security policy
ABOUT DERIV
For over 26 years, Deriv has been committed to making online trading accessible to anyone, anywhere. Trusted by over 3 million traders worldwide, the company offers an expansive range of trade types across popular markets on its award-winning trading platforms. With a workforce of more than 900 people across 13 countries globally, Deriv continues to invest in AI literacy at every level and in building a workplace where its people can grow alongside the business. As the company drives an AI-first transformation across every function, its values of integrity, competence, teamwork and customer focus remain the constant.
PRESS CONTACT pr@deriv.com
