Agentic Gateway

Central middleware routing all LLM and MCP requests through authenticated keys. One endpoint for every AI interaction — unified cost tracking, policy enforcement, and audit trail.

Quick Start

Create a Gateway Key

Go to your workspace Settings > Gateway and generate a key.

Make an LLM call with the OpenAI SDK

No custom dependencies — use the standard OpenAI SDK and point it at Dobby.

First, install the SDK: pip install openai

Python

from openai import OpenAI

client = OpenAI(
    api_key="<YOUR_GATEWAY_KEY>",
    base_url="https://dobby-ai.com/api/v1/gateway"
)

response = client.chat.completions.create(
    model="claude-sonnet-4-20250514",
    messages=[{"role": "user", "content": "Hello from Dobby!"}]
)

print(response.choices[0].message.content)

Uses the standard OpenAI SDK — no custom dependencies. Switch providers by changing the model parameter. Supports Claude, GPT-4, Gemini, Mistral, and 9 more.

See your cost

Open the Gateway Usage tab in your dashboard. Every request is tracked with cost, latency, and token count.

API Key Tiers

Prefix	Type	Rate Limit	Use Case
gk_user_*	User	100 RPM	For individual developers
gk_svc_*	Service	500 RPM	For backend services and agents
gk_tmp_*	Temporary	50 RPM	Auto-expire, for testing or demos

Capabilities

LLM Proxy

OpenAI-compatible endpoint for 13+ providers — Claude, GPT-4, Gemini, Mistral, and more. Switch models by changing one parameter.

MCP Server

73+ tools via JSON-RPC 2.0. Connect Claude Desktop, Cursor, VS Code, or any MCP-compatible client.

Cost Tracking

Automatic per-request cost attribution by agent, user, and provider. Budget enforcement with alerts at 80/90/100%.

Policy Engine

Model restrictions, content filtering, DLP (9 PII patterns), rate limiting, and 4-layer policy hierarchy.

Agent Proxy

Route requests from external agents (CrewAI, LangSmith, custom REST). Auto-discovery registers agents on first call.

Kill-Switch

Emergency control to block all gateway traffic for an organization. Three scopes: all, LLM-only, new keys only.

Key Endpoints

POST /api/v1/gateway/chat/completions— LLM proxy (OpenAI-compatible)

POST /api/v1/gateway/mcp/endpoint— MCP tools (JSON-RPC 2.0)

GET /api/v1/gateway/models— List available models

GET /api/v1/gateway/health— Public health check

Full API Reference Gateway Academy Guide