Opengeni - Agents in your product, infrastructure out of the box

Opengeni is open-source infrastructure for embedding AI agents in products, with durable sessions, sandboxes, credentials, tools, memory, multi-tenancy, and React components.

Visit Website
Opengeni - Agents in your product, infrastructure out of the box

Introduction

Overview

Opengeni is open-source infrastructure for adding AI agents to a product without building the supporting platform from scratch. Teams can use the managed cloud or operate the Apache 2.0 software in their own infrastructure.

The same API supports the Opengeni app and custom product experiences, helping teams embed an agent under their own brand while keeping backend credentials out of the browser.

Key Features

  • Durable sessions continue through dropped connections or worker restarts and can resume from recorded events.
  • Isolated sandboxes separate agent execution from application secrets.
  • Credential handling supports per-user connections and scoped, short-lived access.
  • Tools and MCP support connect APIs that agents need to perform work.
  • Memory preserves relevant workspace and user information across sessions.
  • Multi-tenancy keeps customer data separated with row-level security.
  • Model flexibility includes supported providers and OpenAI-compatible endpoints.

Product Integration

Opengeni provides React components for session conversations, streaming responses, tool steps, and message composition. Styling variables control elements such as accent color, corner shape, typography, and light or dark themes.

A product backend proxies session routes and retains the API key, while the browser communicates through the product's own endpoint. This allows the same agent session to appear in Opengeni's app or inside a custom interface.

Deployment Options

The managed Opengeni Cloud includes hosted infrastructure, updates, sandboxed execution, audit controls, and usage controls. Self-hosting is supported on Kubernetes, with Terraform options for AWS, Azure, and GCP. Enterprise deployments can be scoped around private infrastructure, integration, security, access, SSO, and commercial support requirements.

Pricing

Self-hosted Opengeni has no software license fee under Apache 2.0, while operators cover their own compute, storage, and model-provider costs. Opengeni Cloud charges the model provider's usage cost plus 5%, with no per-seat fee, platform fee, or minimum commitment. Enterprise scope, services, and terms are agreed with each organization.