TengineAI
Login

Introduction

What is TengineAI?

TengineAI is a hosted MCP execution layer for authenticated tool calls.

It provides a remote MCP server that exposes authenticated tools to AI runtimes. Models request tools, TengineAI handles authentication and execution, and structured results are returned to the model.

TengineAI does not provide a client SDK. Instead, it integrates with existing MCP-capable model SDKs and clients, acting as a remote execution layer for authenticated tool calls.

TengineAI works with any MCP-compatible model SDK or client. Anthropic is used throughout the documentation as the reference implementation because it is the most widely used and fully tested today.

How TengineAI Works

┌─────────────┐         ┌──────────────┐         ┌─────────────────┐
│             │         │              │         │                 │
│  MCP Client │────────▶│  TengineAI   │────────▶│  External APIs  │
│  (Runtime)  │         │  (Execution) │         │  (Gmail, etc.)  │
│             │◀────────│              │◀────────│                 │
└─────────────┘         └──────────────┘         └─────────────────┘

Flow:

  1. An MCP client (Claude Desktop, Cursor, or programmatic SDK) connects to TengineAI
  2. The client requests available tools
  3. TengineAI returns the project's enabled tools
  4. The model selects a tool to execute
  5. TengineAI authenticates with the external API using stored credentials (OAuth tokens or API keys)
  6. TengineAI executes the tool call and returns results
  7. The model receives the results and continues execution

Core Concepts

TengineAI uses four primitives:

Projects

Isolated execution environments. Each project has its own API keys, integrations, and enabled tools.

Tools

Actions the model can execute. Examples: gmail-send_email, reddit-get_trending, google-calendar-create_event.

Tools are grouped by provider and require either OAuth or API key authentication.

Integrations

MCP client connections. An integration defines how a model runtime (Claude Desktop, Cursor, custom SDK) authenticates with TengineAI.

Integrations do not define behavior - only connectivity.

API Keys

Project-scoped credentials. API keys authenticate requests from MCP clients to TengineAI. They do not authenticate with external APIs - TengineAI handles that separately.

API keys authenticate MCP clients to TengineAI, not models to providers.

Why Remote Execution Exists

Running tools remotely through TengineAI provides three operational advantages:

1. Centralized Authentication

OAuth tokens and API credentials are stored server-side. Models never see secrets. Credentials can be rotated without redeploying clients.

2. Tool Rotation

Enable or disable tools without changing client configuration. Add new providers without updating installed applications.

3. Usage Control

Enforce rate limits, log all tool calls, and prevent runaway execution. Projects have configurable daily and monthly limits.

What TengineAI Is Not

TengineAI is not:

  • An agent framework
  • A model hosting service
  • A workflow orchestrator
  • A prompt management system

It is execution infrastructure. Models decide what to call. TengineAI executes it safely.

Next Steps