OpenLivery: self-hosting multi-tenant WhatsApp AI agents for agencies

Description and use

OpenLivery lets agencies manage multiple clients from a single self‑hosted deployment, providing isolated workspaces, knowledge bases, and branded portals.

Problem

Agencies need a single installation to manage multiple clients’ WhatsApp AI agents, each with its own knowledge base, conversations, and branded portal, while retaining control over code and integrations.

Outcome

A single installation enables agencies to manage multiple client workspaces, each with its own knowledge base and branded portal, while keeping full control over code, integrations, and costs.

Workflow

• Clone the repo and navigate into it
• Run ./scripts/generate-docker-env.sh to create environment variables
• Execute make up to start Docker containers
• Configure model provider and messaging channel via prompts and README instructions
• Create client workspaces and upload distinct knowledge bases
• Test conversations in each workspace and use human takeover
• Add HTTP tools or MCP servers for agent capabilities
• Iterate and refine based on test results

Evidence

The repository contains the application, rather than just an agent prompt or workflow template: ... Next.js frontend and a FastAPI backend.

Source

With Git, Docker Compose and make available, the repository’s quick start is: ... git clone ... make up

Source