Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/arrozet/caret/llms.txt

Use this file to discover all available pages before exploring further.

Caret is an agentic, AI-first document editor that fills the gap between powerful code AI tools (like Cursor and Copilot) and the stagnant state of document writing. While Word’s Copilot and Google Docs’ Gemini offer chat-toggle AI with minimal editor integration, Caret weaves AI directly into the editing experience — inline ghost-text suggestions, streaming rewrites, and workspace-aware context via RAG.

Quick Start

Run Caret locally with Docker Compose in minutes and open your first document.

Core Features

Explore the rich editor, AI assistant, real-time collaboration, and workspace management.

Architecture

Understand how the microservices, database, and protocols fit together.

API Reference

Full REST and WebSocket API documentation with request/response schemas.

What Makes Caret Different

Caret is built on the conviction that document AI should be native to the editor, not bolted on as a chat sidebar. Every AI capability — streaming text generation, inline suggestions, context retrieval — flows through Tiptap transactions and Y.js to keep collaborative state consistent.

Agentic AI

General and analyst agents with SSE streaming and multi-provider LLM support (OpenAI, Anthropic, OpenRouter).

Real-Time Collaboration

Y.js-powered sync with live cursors, presence awareness, and role-based access control.

RAG Context

pgvector embeddings give the AI workspace-scoped document context for grounded suggestions.

Getting Started

1

Clone the repository

Clone arrozet/caret and copy the example environment file.
2

Configure environment variables

Add your Supabase credentials, database URL, and LLM provider keys to .env at the repo root.
3

Start all services

Run docker compose up --build to start the frontend, API Gateway, document service, collaboration service, and AI service.
4

Open the app

Navigate to http://localhost:5173, sign in with Supabase Auth, and start writing.
Caret has no local PostgreSQL — all services connect to Supabase Cloud via DATABASE_URL. You need a Supabase project before running locally.

Build docs developers (and LLMs) love