IRIS Developer Platform

The Agent OS for Developers

50+
Built-in Tools
300+
AI Models
250+
Integrations
<5min
Setup Time
Why Build on IRIS

Everything you need to ship AI agents

Stop stitching together LangChain + Pinecone + Stripe + Auth + CRM. IRIS is the full stack — from agent creation to revenue collection.

Agent OS

Complete runtime for AI agents. Multi-model support (GPT-4o, Claude, Gemini, Llama, Ollama), tool calling, persistent memory, and autonomous execution.

50+ Built-in Tools

YAML-driven tool registry with web search, RAG, lead gen, file system, CRM, publishing, analytics — all health-checked and ready.

Workflow Engine

Visual drag-and-drop builder with multi-agent coordination, human-in-the-loop, event-driven automation, and real-time progress monitoring.

Memory & Context

Persistent memory vault with Pinecone RAG, automatic file vectorization, user-isolated knowledge bases, and two-way sync.

Repos as Knowledge Bases

Clone any GitHub repo and turn it into a queryable agent. Automatic indexing, semantic code search, and integration with Claude Code, Cursor, and OpenCode.

Agent Marketplace

Publish agents as services. Discovery feed, trending templates, skills library. Users install and subscribe — you earn.

Monetization Built In

Stripe subscriptions, usage-based billing, payment retry, tier management — charge for your agents from day one.

Workspace UI & Workflow Builder

Visual drag-and-drop canvas for building multi-step agentic workflows — no code required.

  • Drag-and-drop workflow canvas with live preview
  • Multi-step agent orchestration with branching logic
  • Human-in-the-loop checkpoints and approval gates
  • Real-time execution monitoring and progress tracking
  • LangGraph (Python) + n8n (visual) dual engine
  • 250+ integrations via Composio, n8n, and built-in connectors
Explore Workflows

Heartbeat & Cron System

Autonomous agent scheduling — agents that work while you sleep.

  • Configurable cron intervals per agent (hourly, daily, weekly)
  • Automatic data gathering and context-aware prompt building
  • Background task execution loop with retry logic
  • Direct database result writing (completed_pending status)
  • Quality evaluation with configurable retry thresholds
  • Pending task carry-forward across heartbeat cycles
Learn About Heartbeats
Component type not implemented: SplitContent
Component type not implemented: CodeShowcase
Component type not implemented: LogoStrip
Agent-Native Platform

Built for Coding Agents

IRIS is natively compatible with every major coding agent. Install our skill and let your AI build workflows for you.

SKILL.md

Teach any coding agent how to use the IRIS platform. Drop our skill file into your project and Claude Code, OpenCode, Cursor, and Codex instantly understand your agent architecture.

AGENTS.md

A README for AI agents. Documents your project structure, API endpoints, and conventions so any coding agent can build extensions and integrations for your IRIS workspace.

/.well-known/skills/

Auto-discovery via web standard (Cloudflare RFC). Any AI agent can find and install your IRIS skills from your domain — like robots.txt for AI capabilities.

Component type not implemented: CodeShowcase

V6 Tool Registry

YAML-driven tool system with 50+ tools. Web search, RAG, file ops, CRM, publishing, analytics — all with health checks and OpenAI function-calling schema.

A2A Bridge

Agent-to-Agent protocol. Connect coding agents (Claude Code, OpenCode), messaging bots (Telegram, Discord), and sessions into a unified dashboard.

Memory Vault

Persistent context with Pinecone vector search. File attachments auto-vectorized. User-isolated knowledge bases with access tokens and two-way sync.

Workflow Engine

Multi-step orchestration with LangGraph (Python) + n8n (visual). Human-in-the-loop, event-driven automation, Redis state, and 250+ integrations.

Agent Marketplace

Publish agents as services. Discovery feed, trending, templates, skills library. Users install and subscribe — you earn. Stripe billing built in.

Multi-Model Gateway

GPT-4o, Claude, Gemini, Llama, Ollama — swap models per-agent or per-request. Intelligent routing, cost optimization, and automatic failover.

For Developers, By Developers
“Stop building AI infrastructure. Start building AI products. IRIS gives you the operating system — you bring the vision.”
Get Started

Build Something Extraordinary

Join the developers shipping AI agents on IRIS. One-line install, full platform, zero infrastructure headaches.