rivenai.io
Models
Every frontier model. One key. One bill.
Frontier, open-weight, and Riven-tuned lanes behind a single OpenAI-compatible endpoint. Pick per message in chat, or pin one in the API.
Explore
Every surface is a path off rivenai.io
Chat
Grounded citations in the answer stream
Every claim carries a numbered link. Follow-up chips write your next prompt.
rivenai.io/chat
One gateway, every model
Riven Auto, Fast, Core, Pro, and Research — pick your model per turn.
rivenai.io/chat
Spaces and Skills you can pass to agents
Load a Space, add a Skill, hand the same context off to /agents.
rivenai.io/chat
Code
Agents
Browse
Enterprise
Riven — Self-hosted, single-tenant AI for enterprises that demand control, compliance, and performance.
Single-tenant by default
Your deployment runs alone: dedicated models, dedicated keys, dedicated data path. Nothing pooled, nothing shared.
Governed by Riven Control
Keys, routing, budgets, and audit in one pane. Every request is metered and attributable — usage math you can hand to finance.
Your infrastructure or ours
Run on Riven-managed capacity, or bring your own cloud and hardware. Same API, same governance, either way.
Apps