RTIE.AI / docs
Home OpenAPI Quickstart
Build with AI

Build with AI on RTIE

RTIE has three AI-facing surfaces: coding-agent context for developers, tenant-scoped MCP tools for operating enabled modules, and runtime agent APIs for agents that join sessions as audited actors.

The AI Tools Suite, prompts, skills, RTIE.md, and LLM-readable docs are available now. Public MCP tool calling at mcp.rtie.ai is in private preview and is tenant/module scoped.

🔌
Agent Plugin

Coding-agent context for building RTIE apps: skills, prompts, commands, and RTIE project rules.

Install plugin →
🛠
MCP Private Preview

Tenant-scoped tool execution for enabled modules. API keys and session tokens resolve to RTIE actors.

Tool reference →
📚
Agent Skills

Procedural instruction sets encoding RTIE-specific knowledge — quickstart, module dev, schema, realtime, auth, and more.

Browse skills →
📋
Prompts

7 copy-paste task prompts for common RTIE workflows — bootstrap, auctions, webhooks, presence, Action Engine, and more.

Browse prompts →

Quick setup

Surface 1 — AI coding tools

Use these when an agent is editing code or reading docs. They do not grant runtime authority by themselves.

claude plugin marketplace add https://docs.rtie.ai/ai-tools/marketplace.json
claude plugin install rtie

This installs RTIE skills, commands, context files, and private-preview MCP configuration. Runtime MCP calls still require tenant credentials.

Surface 2 — MCP tenant tools

Add this only for tenants with MCP private-preview access. Tool availability is filtered by workspace type, plan, enabled modules, and key/session scope.

{
  "mcpServers": {
    "rtie": {
      "type": "http",
      "url": "https://mcp.rtie.ai/mcp",
      "headers": { "Authorization": "Bearer ${RTIE_API_KEY}" }
    }
  }
}

Surface 3 — Runtime agents

Runtime agents join rooms or sessions through /v1/runtime/connect and act with short-lived module/session tokens. Use human confirmation for bids, buys, payments, destructive config, and rule replacement.

Cursor / other agents

Download .mcp.json or rtie.mdc and add to your project.


What's included

Component What it is Distribution
MCP private preview Planned typed tools, tenant-scoped, manifest-driven mcp.rtie.ai
Plugin Claude Code plugin with private-preview MCP config + skills + commands docs.rtie.ai/ai-tools
Skills (9) Procedural instruction sets for AI agents In plugin
Prompts (7) Copy-paste task prompts docs.rtie.ai/ai-tools/prompts
RTIE.md AI coding context file rtie.ai/RTIE.md
llms.txt LLM-readable site index rtie.ai/llms.txt
llms-full.txt Full developer docs for LLMs docs.rtie.ai/llms-full.txt

Need an API key? cloud.rtie.ai → API Keys → Create key.