Introduction

What is Riven AI?

Riven AI is an AI company building the full stack — from training and serving our own models to orchestrating autonomous agents and managing production infrastructure. We own every layer, on a path toward AGI and robotics.

Whether you are training custom models, deploying autonomous agents, building internal tools, or managing production infrastructure, Riven AI provides the tools and abstractions to ship faster with confidence.

Core Products

The platform is built around six core products, each addressing a critical part of the AI application lifecycle:

• AI Platform — Train, evaluate, and serve AI models including LLMs, embedding models, and image generation pipelines. Supports vLLM, custom model registries, and automatic scaling.
• Agents — Build and orchestrate autonomous AI agents equipped with skills, persistent memory, and tool-use capabilities. Agents can operate independently or collaborate on complex workflows.
• Knowledge Engine — Crawl, ingest, and semantically search across your organization's knowledge. RAG pipelines with automatic chunking, embedding, and corpus management.
• App Builder — Build applications with AI assistance. Natural language to working code with a full editor, live preview, and one-click deployment.
• Dev Center — Service catalog, deployment management, and API documentation for your microservices. Proto-first API design with Connect RPC transport.
• Infrastructure — Production-grade Kubernetes orchestration, Helm-based deployments, CI/CD pipelines, and full observability with Grafana, Prometheus, and Jaeger.

Architecture Overview

Riven AI follows a modular architecture where each product operates as an independent subsystem but shares common infrastructure, authentication, and observability layers.

┌─────────────────────────────────────────────────────────────┐
│                     Riven Platform UI                       │
├──────────┬──────────┬──────────┬──────────┬────────┬────────┤
│    AI    │  Agents  │Knowledge │   App    │  Dev   │ Infra  │
│ Platform │  Manager │  Engine  │ Builder  │ Center │ (K8s)  │
├──────────┴──────────┴──────────┴──────────┴────────┴────────┤
│            Auth  ·  Observability  ·  Storage                │
├─────────────────────────────────────────────────────────────┤
│                    Kubernetes (EKS)                          │
└─────────────────────────────────────────────────────────────┘

All services communicate via Connect RPC (gRPC-Web compatible) and are deployed as Helm-managed workloads on Amazon EKS. Authentication is handled centrally through JWT/JWKS tokens and OpenFGA-based authorization.

Next Steps

Ready to get started? Follow these guides to set up your environment and deploy your first project:

• Installation — Detailed setup instructions for the CLI and SDKs.
• Quick Start — Get up and running in under 5 minutes.
• Configuration — Configure your environment and project settings.
• Authentication — Set up authentication for the CLI and CI pipelines.