Documentation

Welcome

Connect your AI to your files, tools, and apps — then offer it to anyone through the channels they already use.

Welcome to the Ganju docs. Ganju connects AI assistants like Claude, ChatGPT, and Gemini to your own files, tools, and apps — no servers to run, no code to write.

Then it lets you offer that AI to whoever you want — clients, your team, or individual people — through the channels they already use: Slack, WhatsApp, Discord, Telegram, and more on the way. One workspace, delivered everywhere.

These guides take you from a blank workspace to a live assistant you can share with anyone.

Start here

• Getting started — create your first workspace and connect your AI in a few minutes.
• Organizations & projects — how Ganju is structured: an organization for your team and billing, a project per assistant.

Explore what Ganju can do

Once you’re set up, these guides go deeper on each capability — the building blocks you combine to shape exactly what your assistant knows, does, and where it lives:

• Prompts — reusable, multi-turn templates with typed variables that become slash commands in every channel.
• Resources — ground your AI in your own knowledge: import from Google Drive or OneDrive, crawl a website, or upload files, all searchable with citations.
• Tools — let it take action: email (Gmail, Outlook), Slack, calendars (Google, Cal.com), web search, remote MCP servers like GitHub and Notion, or your own HTTP APIs.
• Channels — deliver it to real people on Telegram, WhatsApp, Slack, and Discord, with step-by-step setup for each.
• MCP clients — connect the same assistant to Claude, ChatGPT, Cursor, or VS Code with one URL.
• Analytics — see activity over time across channels and clients, resource/tool/prompt usage, and a live recent-activity feed.
• Settings — manage your plan and billing, invite your team, and bring your own language model.
• Deploy it yourself — Ganju is open source (Apache-2.0); run your own instance on Cloudflare and Postgres, with no plan caps.

Every page on this site is also available as Markdown — just add .md to the URL (you’re reading /docs/welcome.md). See /llms.txt for the full index.