Skip to main content
Product · Open Source

CrewMeld

Enterprise AI digital workforce platform — manage AI employees like real team members. Open-source and privately deployable, with multi-model, multi-channel and visual workflow orchestration.

Overview

Turn AI capability into your company’s digital workforce

CrewMeld is an open-source, enterprise-grade AI digital workforce platform. It packages the power of large models and agents into manageable, collaborative and auditable "digital employees" — letting AI take on clear responsibilities, follow defined processes and work seamlessly alongside people, instead of staying stuck in a one-off chat box.

A visual SOP editor sits at its core, supporting drag-and-drop orchestration of multi-step, multi-role workflows with conditional branches and human approval steps. Combined with a knowledge base and RAG retrieval, every answer and decision a digital employee makes is grounded and traceable. Tools run isolated in sandboxes, while three-tier RBAC and full audit logging meet enterprise security and compliance needs.

For connectivity, CrewMeld supports 13+ model providers including OpenAI, Anthropic, Qwen, DeepSeek and Ollama — switchable by cost, latency and scenario — and integrates 8+ channels such as WeCom, DingTalk, Lark, Telegram, Discord, email and SMS, putting digital employees right where teams already work.

For delivery, CrewMeld runs on Docker Compose, Kubernetes / Helm or fully on-premise private deployment, keeping data in your domain. As an open-source project, enterprises can adopt it for free and self-serve, build their own agent applications on top of it, or have us provide deep customization and private rollout support.

Product tour

Real interfaces — from the agent workspace and digital-employee roster to multi-channel integration.

Key features

Visual SOP editor

Drag-and-drop multi-step collaborative workflows with conditional logic.

13+ LLM providers

OpenAI, Anthropic, Qwen, DeepSeek, Ollama and more.

Multi-channel integration

WeCom, DingTalk, Lark, Telegram, Discord, email, SMS.

Knowledge base

Document management with hybrid vector / keyword search (RAG).

RBAC & audit

Three-tier role management and compliance audit logs.

Flexible deployment

Docker Compose, Kubernetes / Helm or on-prem.

Tech stack

Next.js 16React 19TypeScriptBunPostgreSQL 17RedisMinIODocker

Use cases

  • Multi-step process automation
  • AI + human approval
  • Knowledge retrieval & Q&A
  • Private / domestic-platform integration

Open-source and ready — build with us

Get CrewMeld on GitHub, or contact us for custom development and private deployment.