Skip to main content
Six chat platforms, one architecture. Each bot maintains per-channel (or per-DM) conversation history, calls the Vaquill statutes API, renders structured statute citations, and enforces per-user rate limits.

Slack

@mentions, slash commands, DMs, thread context, feedback buttons. Socket Mode or HTTP.

Discord

Prefix commands, per-channel history, pagination, starter questions, role and channel allowlists.

Microsoft Teams

Adaptive Cards, thread support, tenant allowlists, Azure-native deployment.

Telegram

BotFather setup, markdown tables rendered as PNG, inline source buttons, polling or webhook.

WhatsApp

Twilio-based, FastAPI server, slash commands for mode/language switching, sandbox to production path.

Signal

signal-cli-rest-api backend, Docker Compose deploy, primary or linked-device registration.

​
Shared architecture

Every chatbot follows the same flow: All six bots share:
  • Vaquill client (vaquill_client.py) that wraps POST /api/v1/statutes/search
  • Conversation manager for per-channel or per-DM chat history
  • Rate limiter with Redis or in-memory backend
  • Slash / prefix commands: help, clear, examples, stats
  • Statute citation rendering appropriate to each platform’s primitives (Block Kit, Adaptive Cards, inline keyboards, etc.)

​
Pick by deployment surface

If you want zero exposed ports during development, Slack (Socket Mode), Discord, Telegram (polling), and Signal all run without a public URL.

​
Deploy targets

All six bots ship with Dockerfiles and have been smoke-tested on:
  • Docker Compose (recommended for self-hosting)
  • Railway (Git-based deploys)
  • Render (free tier with cold starts)
  • Fly.io (low-latency regions)
  • VPS + systemd (full control)
  • Azure App Service (Teams bot only)
Each bot’s page lists its specific deployment recipe.

​
Source

All chatbot code lives at github.com/Vaquill-AI/integrations under one subdirectory per platform. Fork the repo, edit one bot, redeploy.