API Reference

The Mako API is a RESTful API built with Hono. All endpoints are prefixed with /api.

Authentication

Method	Endpoint	Description
POST	/api/auth/login	Login with email/password
POST	/api/auth/register	Create a new account
GET	/api/auth/me	Get current user session
POST	/api/auth/logout	End session

Workspaces

Method	Endpoint	Description
GET	/api/workspaces	List user workspaces
POST	/api/workspaces	Create a workspace
GET	/api/workspaces/:id	Get workspace details

Connectors & Flows

Method	Endpoint	Description
GET	/api/workspaces/:wid/connectors	List configured data sources
POST	/api/workspaces/:wid/connectors	Add a new data source
POST	/api/workspaces/:wid/flows	Trigger a flow
GET	/api/workspaces/:wid/flows/:fid	Get flow status

Query Execution

Method	Endpoint	Description
POST	/api/execute	Execute a SQL/NoSQL query

AI Agent / Chat API

The Chat API enables programmatic access to Mako’s AI agent. It supports both session-based and API key authentication.

Method	Endpoint	Description
POST	/api/agent/chat	Send messages and receive streaming AI response
GET	/api/agent/models	List available AI models
GET	/api/agent/agents	List available agent modes

Authentication

Use an API key in the Authorization header:

Authorization: Bearer revops_YOUR_API_KEY

API keys are created in Settings → API Keys. Each key is scoped to a workspace. Chats created via API key are attributed to the key’s creator, so they appear in the creator’s chat history.

Request Body (POST /api/agent/chat)

Field	Type	Required	Description
messages	UIMessage[]	Yes	Array of messages. Each has id, role, and parts.
chatId	string	Yes	24-char MongoDB ObjectId. New ID for new chats; reuse for follow-ups.
workspaceId	string	Yes	24-char MongoDB ObjectId. Must match the API key’s workspace.
modelId	string	No	Model ID (e.g. gpt-5.2, gemini-2.5-flash). Default: gpt-5.2.
openConsoles	object[]	No	Console context for SQL tools.
consoleId	string	No	Active console ID.
agentId	string	No	Agent mode: console (default) or flow.

Message Format

{
  "id": "unique-id",
  "role": "user",
  "parts": [{ "type": "text", "text": "Your message here" }]
}

Example Request

curl -X POST https://your-mako.com/api/agent/chat \
  -H "Authorization: Bearer revops_YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "messages": [{
      "id": "1",
      "role": "user",
      "parts": [{ "type": "text", "text": "Show top 5 customers" }]
    }],
    "chatId": "507f1f77bcf86cd799439011",
    "workspaceId": "YOUR_WORKSPACE_ID",
    "modelId": "gemini-2.5-flash"
  }'

Response Format

The response is a streaming SSE (Server-Sent Events) stream. Key event types:

Event Type	Description
start	Stream started. Contains messageId.
text-delta	Incremental text chunk from the assistant.
tool-input-available	Tool call with full input (e.g. SQL query).
tool-result	Tool execution result (e.g. query results).
finish	Stream finished. Contains finishReason.

Error Codes

Status	Description
400	Missing or invalid messages, chatId, or workspaceId.
401	Unauthorized (missing or invalid API key).
403	API key not authorized for the workspace.
404	Agent not found (invalid agentId).

For full details, see CHAT_API_REFERENCE.md.

Chat History

Method	Endpoint	Description
GET	/api/workspaces/:wid/chats	List chat sessions
POST	/api/workspaces/:wid/chats	Create a new chat
GET	/api/workspaces/:wid/chats/:id	Get chat details
PUT	/api/workspaces/:wid/chats/:id	Update chat title
DELETE	/api/workspaces/:wid/chats/:id	Delete a chat

Console API

For executing saved consoles programmatically, see CONSOLE_API_DOCUMENTATION.md.

Inngest

Method	Endpoint	Description
POST	/api/inngest	Inngest webhook handler