DeskcommCRM

DeskcommCRM is an open‑source, self‑hosted CRM that adds AI agents to WhatsApp conversations. It aims to keep every interaction recorded and auditable while running on your own server without per‑user subscription fees.

The platform combines a traditional lead‑management interface with AI‑driven agents that can qualify contacts, move them through the sales funnel, and trigger follow‑up actions. It is built for businesses that sell primarily through chat and need a system that respects Brazilian data‑protection rules (LGPD). Because the code is released under an MIT license, you can inspect the architecture, modify it, and deploy it on a VPS that meets the minimum recommendation of 4 GB RAM.

To get started you clone the GitHub repository, run a single install script, and the system pulls all required containers, generates secrets, creates the database, and starts the services. After installation you access the web UI, create an administrator account, and scan the WhatsApp QR code to link a phone. The platform automatically provisions an HTTPS certificate on first access, so you do not have to configure TLS manually. Updates are applied with another script that backs up the database, reapplies the schema idempotently, and runs health checks, all documented in the repository.

When a WhatsApp message arrives, the AI agent first looks up the contact's history, then consults your knowledge base for policies, catalog items, or delivery zones. Before sending any reply, it runs seven checks, such as LGPD compliance, anti‑spam rules, and deterministic promise validation, and records both the attempted and the blocked messages. If the lead is qualified, the system moves it to the next stage, adds tags, and assigns an owner, all with a visible audit trail. Human agents receive a summarized context of the conversation, and if no reply is sent, the lead is placed in a “Radar” view where it is classified by risk and scheduled for follow‑up, preventing it from disappearing unnoticed.

Key features

• AI agents handle WhatsApp messages, qualify leads and update the sales funnel automatically
• All interactions and AI decisions are logged and can be audited through a timeline view
• Built‑in LGPD compliance checks and anti‑spam safeguards before each outbound message
• Self‑hosted deployment with a single‑command install script and automatic HTTPS provisioning
• Multi‑tenant architecture and public documentation of the system's five core rules