DeskcommCRM: An Open-Source AI Sales CRM for WhatsApp You Can Self-Host on a Single VPS

Forum topic · ✨步子哥 · 2026-09-12

Summary

DeskcommCRM is an open-source, self-hosted AI sales platform that gained 505 GitHub stars in a single day. Built by a Brazilian developer as a self-hosted alternative to Kommo, Octadesk, and Intercom, it bundles Next.js 16, Supabase, a WAHA WhatsApp gateway, and an AI agent into a one-command install on a 4GB VPS. The installation script validates inputs, is idempotent, and adapts to existing reverse proxies, making deployment feasible for non-technical users. Unlike typical CRM chatbot features, its AI agent actively handles customers, scores and routes leads using rule-based logic, auto-replies to low-value contacts, and escalates high-value leads to humans. The project is also 'MCP-ready': Claude Code, Codex, Cursor, and other coding agents can read bundled skill files and perform the entire installation. Targeting WhatsApp-first markets like Brazil, India, and Indonesia, DeskcommCRM exemplifies AI agents as process executors rather than smarter chat boxes.

A Brazilian developer has open-sourced DeskcommCRM — a project that squeezes an entire sales team onto a single VPS, using WhatsApp as the battlefield and an AI agent to greet customers, filter leads, and close deals. The project gained 505 GitHub stars in a single day. Its positioning is straightforward: an open-source AI sales OS, a self-hosted alternative to Kommo, Octadesk, and Intercom.

One Command, One CRM

The pain point of traditional SaaS CRMs isn't a lack of features — it's that you pay for someone else's servers every month. DeskcommCRM flips that logic: rent a 4GB VPS, run one bash hostgator-setup-kit/install.sh, and the entire stack (Next.js 16 + Supabase + WhatsApp gateway + AI agent) comes up. HTTPS is auto-configured, the database schema is auto-created, the admin user is auto-provisioned, and even cron jobs are installed automatically.

This isn't just "one-click deploy" marketing. The install script does three things traditional one-click scripts don't:

1. Validates every input — a wrong API key is rejected immediately, not three steps later 2. Idempotent — rerunning it breaks nothing and doesn't duplicate cron jobs 3. Environment-adaptive — if the VPS already has a reverse proxy (Hostinger/Coolify), it switches publishing modes instead of blindly spinning up another Caddy

Together, these mean non-technical users can install it — solving a long-tail problem open-source CRMs never fixed.

The AI Agent Isn't a Gimmick, It's the Core

Most CRMs' "AI features" are a chat box for asking "what were this month's sales?" DeskcommCRM's AI agent does something else: greeting customers, filtering leads, and closing deals.

On WhatsApp, the agent does three things:

Reception: catches incoming customer messages, understands intent, and judges whether it's a valid lead
Filtering: scores, tags, and routes leads to queues based on preset rules (WHEN/IF/THEN)
Sales: auto-replies to low-value leads and escalates high-value ones to humans

This isn't an "AI assistant" — it's an "AI salesperson." An assistant answers your questions; a salesperson does the work for you.

MCP-Ready: Let Claude Code Install Your CRM

The most interesting part of the README:

> Drop the hostgator-setup-kit/ folder into the chat of Claude Code running inside the VPS and say "install DeskcommCRM for me."

Hand the install kit to Claude Code, which reads the installation guide in CLAUDE.md and walks you through deployment in Portuguese. This isn't an "AI feature" — it's using an AI agent as your ops team.

The project also ships a .agents/skills/deskcomm-instalar/ directory that Claude Code, Codex, Cursor, OpenCode, and Antigravity auto-load when opening the project folder. It's "skill-driven deployment" — instead of writing prompts for AI, you write skills for AI.

Why WhatsApp

DeskcommCRM was written by a Brazilian, and Brazil is WhatsApp's biggest market. Brazilian small businesses don't use email, Slack, or DingTalk — all business happens on WhatsApp. So a "WhatsApp-native CRM" isn't a feature in Brazil; it's a survival requirement.

But the positioning holds globally: India, Indonesia, Mexico, Nigeria — SMBs across emerging markets run their business on WhatsApp. DeskcommCRM targets not Silicon Valley's SaaS market but the WhatsApp sales scenario of 500 million SMBs worldwide.

Tech Stack Rationale

Next.js 16: latest version, App Router + Server Actions, unified front and back end
Supabase: Postgres + Auth + Storage, eliminating three services
WAHA (WhatsApp HTTP API): open-source WhatsApp gateway, avoiding Meta's official API (expensive and slow to approve)
TypeScript strict: mandatory, not optional

The core judgment behind this stack: self-hosting must be simple. Supabase removes three services, WAHA skips Meta's review process, Next.js avoids split frontend/backend deployment. The whole stack needs only one VPS, a free Supabase account, and one AI API key.

Worth Watching

DeskcommCRM isn't the most technically innovative project, but it gets several things right:

1. Puts the AI agent in the right place — not a chat box, but the execution layer of the sales pipeline 2. Minimizes deployment friction — non-technical users can install it, the long-tail pain of open-source CRMs 3. Treats skills as a deployment interface — Claude Code reads skill files and installs for you, an early form of "AI ops" 4. Picked the right market — WhatsApp sales is a must-have in emerging markets, not a nice-to-have

505 stars isn't huge, but the project's significance isn't in the star count. It proves one thing: the landing scenario for AI agents isn't "a smarter chat box" — it's "a process executor that works on your behalf." Sales CRM is the first vertical being rewritten by AI agents, and DeskcommCRM is a concrete sample of that trend.

---

GitHub: https://github.com/melgarafael/DeskcommCRM