Rei Chat quickstart

Authenticate, choose the correct Rei Chat base URL, and make your first API call.

The Rei Chat platform exposes two complementary HTTP APIs: the Agent API for managing your agents and their configuration, and the Inference API for running chat and response generation against an agent. This guide explains the scope of each, which credential to use, and how to make your first request.

APIs at a glance

API	Base URL	Scope	Auth
Agent API	https://api.reilabs.org/v1	Create, configure, and manage agents (units) and their tools	Account API Key and Unit API Key (per endpoint)
Inference API	https://rei-api.reilabs.org/v1	Run chat completions and generate responses from an agent	Unit API Key

All paths are versioned under /v1. Always send your credential in the Authorization header as a bearer token: Authorization: Bearer <SECRET_KEY>.

---

Agent API

Base URL: https://api.reilabs.org/v1

Scope & usage: Use the Agent API to manage the lifecycle and configuration of your agents: the control-plane operations you typically run from a backend or during setup. It is split into two sub-categories based on which credential authorizes the call.

Account & Unit Management

Operations that create or change agents on your account. These are authorized with your Account API Key, because they act across your account rather than on behalf of a single agent.

Method	Endpoint	Description
POST	/accounts/units	Create a new agent (unit)
PUT	/accounts/units	Update an existing agent
DELETE	/accounts/units	Delete an agent
POST	/accounts/units/clone	Clone an existing agent

Agent Operations

Operations that read or update a single agent and its tools. These are authorized with that agent’s Unit API Key.

Method	Endpoint	Description
GET	/agents	Retrieve the agent’s details
GET	/agents/tools	List the agent’s tools
PUT	/agents/tools	Update the agent’s tools

---

Inference API

Base URL: https://rei-api.reilabs.org/v1

Scope & usage: Use the Inference API at runtime to interact with a configured agent, sending messages and receiving generated output. All Inference API calls are authorized with the Unit API Key of the agent you are calling.

Chat & Responses

Method	Endpoint	Description
POST	/chat-completion	Run a chat completion against the agent
POST	/response	Generate a response from the agent

---

Authentication

Rei Chat uses two types of bearer credentials. Pick the one required by the endpoint you are calling (see the tables above).

Credential	Used for	Endpoints
Account API Key	Managing resources across your account	POST/PUT/DELETE /accounts/units, POST /accounts/units/clone
Unit API Key	Acting as a specific agent	GET /agents, GET/PUT /agents/tools, POST /chat-completion, POST /response

Account API Key

To manage resources.

Unit API Key

To act on behalf of a specific agent.

---

Make your first request

Once you have an Unit API Key, call the Inference API to run a chat completion:

curl -X POST https://rei-api.reilabs.org/v1/chat-completion \
  -H "Authorization: Bearer YOUR_UNIT_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "messages": [
      { "role": "user", "content": "Hello, who are you?" }
    ]
  }'

To manage agents instead, call the Agent API with your Account API Key:

curl -X POST https://api.reilabs.org/v1/accounts/units \
  -H "Authorization: Bearer YOUR_ACCOUNT_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "agentModel": "google/gemini-2.5-flash",
    "tag": "my-first-agent",
    "behaviourPrompt": "You are a helpful assistant."
  }'

See the Agent API and Inference API sections in the sidebar for the full request and response schema of every endpoint.