Belong AI Deploy

Overview

Belong AI Deploy sets up a private AI agent that runs on your own cloud account. Connect it to Telegram, WhatsApp, Slack, or another supported channel, then operate it through a Tailscale-private dashboard. You keep the infrastructure, data, and AI-provider relationship under your control.

Choose the path that fits how you work:

• Guided hosted setup: Launch at aideploy.co with provider sign-ins, a visual deployment flow, and managed operations.
• Open-source self-hosting: Explore belongnet/aideploy on GitHub for the public CLI, runtime stack, OpenTofu modules, lifecycle tools, and source.

The guided hosted deploy

1. Sign in with DigitalOcean — one click, your server is ready
2. Sign in with ChatGPT — one click, your AI is connected
3. Connect Telegram — paste one token
4. Click Launch — your agent is live in about 60 seconds

See the Quick Start for the step-by-step walkthrough.

How it works

subgraph You[Your devices]
  D[Phone / Laptop<br/>Tailscale]
end

subgraph Cloud[Your cloud server]
  DASH[Dashboard]
  AG[AI Agent]
  GW[Gateway]
  DB[(Private memory DB)]
  AG --- GW
  AG --- DB
  DASH --- DB
end

M[Telegram / WhatsApp / Slack]

D -- private network --> DASH
M -- optional webhooks --> GW

style You fill:#e1f5fe
style Cloud fill:#f3e8ff

• Your dashboard and agent are private — reachable only through your own Tailscale network, not the public internet.
• Messaging webhooks are optional and, when enabled, only open a single HTTPS route for the channel you chose.
• Your agent stores long-term memory in its own private database, so it remembers context across conversations.

Hosted support

Cloud

See Cloud Providers for details.

AI

Image generation and voice-note transcription ship preconfigured. See AI Providers.

Agent runtimes

Both runtimes are available in the public repository and the hosted wizard. Their channel support differs; the wizard shows only compatible combinations.

Messaging

See Channels.

What you can do after deploy

Once your agent is live, open the management dashboard over your private network to:

• Chat & review — read every conversation in a threaded view, search, and star important ones
• Automate with tasks — trigger actions on keywords, schedules, or webhooks (welcome greetings, FAQ auto-replies, daily summaries, and more)
• Tune your agent — set its personality/system prompt, switch models, and manage conversation cleanup
• Run more agents — deploy multiple independent agents on one server, each with its own model and channels

Server sizes & cost

Most people should start with Starter — you can resize later from your cloud provider's dashboard. Your AI usage is billed by your provider (or covered by your existing subscription when you use OAuth).

Deploying programmatically

Teams that want automation instead of the sign-in wizard can provision agents with a REST API or a local MCP server (deploy from CI, manage a fleet from your editor, or batch-launch an agent per customer). See Enterprise & API.

Next steps

• Quick Start — deploy your first agent
• Open Source — self-host from the public repository
• Cloud Providers — choose where it runs
• AI Providers — choose its brain
• Private Access — reach your dashboard securely
• Enterprise & API — programmatic deployment via REST API or MCP
• Troubleshooting — fix common issues