Learn to Build with Clawix

Clawix is an open-source, self-hosted multi-agent AI orchestration platform designed for secure AI workflow execution, governance, memory management, and scalable agent collaboration inside isolated Docker containers.

Clawix is designed for:

• AI consultants building enterprise AI workflows
• AI enthusiasts experimenting with autonomous agents
• Junior developers learning AI orchestration and infrastructure
• Organizations requiring private and governable AI systems

Clawix focuses on production-grade orchestration with:

• Container-isolated agents
• Self-hosted deployment
• Governance and auditability
• Multi-agent collaboration
• Multi-provider AI support
• Persistent memory and workspaces
• Role-based access control

Yes. Clawix is publicly available as an open-source project under the MIT license.

A multi-agent orchestration platform enables multiple AI agents to collaborate, delegate tasks, share memory, and execute workflows together.

AI agents are autonomous assistants powered by large language models (LLMs) that can execute tasks, use tools, and coordinate workflows.

Sub-agents are specialized helper agents dynamically created by primary agents to perform delegated tasks independently inside isolated containers.

Swarm orchestration allows multiple AI agents to collaborate in Directed Acyclic Graphs (DAGs), where tasks can run sequentially or in parallel before results are aggregated into a final response.

Heartbeats and cron workflows allow AI agents to execute recurring scheduled tasks such as reminders, monitoring, automation jobs, and periodic reporting.

Minimum requirements:

• Node.js 20+
• pnpm 9+
• Docker
• Docker Desktop

Recommended:

• Apple Silicon Mac
• 16GB RAM or above

Yes. Docker is required because every AI agent runs inside its own isolated container for security, sandboxing, and resource management.

Installation steps:

1. Clone the repository
2. Start Docker Desktop
3. Run: pnpm run install:clawix

The installer automatically configures environment variables, builds Docker images, and starts required services.

No. The interactive installer automatically generates the .env file and secure credentials. Manual editing is optional for advanced customization.

The installer:

• Checks prerequisites
• Generates secure secrets
• Configures AI providers
• Creates the initial admin account
• Builds Docker images
• Starts PostgreSQL and Redis
• Boots the Clawix platform

Run: pnpm run dev

Then open: http://localhost:3000

Common deployment commands:

• pnpm run install:clawix
• pnpm run update:clawix
• pnpm run docker:dev
• pnpm run docker:down

Yes. Clawix is fully self-hosted, allowing organizations to retain ownership of infrastructure, models, and data.

Yes. Clawix is designed for production-grade deployment with isolated execution, governance controls, persistent storage, and multi-user support.

Recommended hardware:

• Apple Silicon Mac Mini
• 16GB RAM or higher
• SSD storage

Clawix can also run on cloud servers and OpenAI-compatible inference infrastructure.

Yes. Lightweight deployments are possible depending on the AI model workload and infrastructure requirements.

Available providers:

• Anthropic Claude
• OpenAI GPT
• Z.AI Coding
• OpenAI-compatible APIs

Planned providers:

• Gemini
• DeepSeek
• Azure OpenAI
• OpenRouter
• Kimi

Yes. Clawix supports OpenAI-compatible endpoints such as Ollama, vLLM, and self-hosted inference APIs.

Yes. Clawix uses a provider abstraction layer that allows providers to be configured through environment settings and provider specifications.

Currently available:

• Web Dashboard
• Telegram

Planned:

• WhatsApp
• Slack

Yes. Telegram integration is fully supported using Telegram bot tokens configured by Admin users.

Yes. Users are identified by their Telegram IDs, allowing multiple users to safely share one Telegram bot.

No. Telegram does not allow the same bot token to be used simultaneously across multiple deployments or running processes.

Yes. Planned WhatsApp and Slack integrations are intended for customer support, enterprise workflows, and business communication.

Container isolation prevents:

• Cross-agent interference
• Unauthorized file access
• Resource exhaustion
• Host-level execution risks

No. Each agent runs inside its own sandboxed Docker container with isolated workspaces and scoped permissions.

Provider API keys and secrets are encrypted using AES-256-GCM encryption and stored securely on the server.

Governance features include:

• Audit logs
• Token budgets
• Structured logging
• Prometheus metrics
• Scoped memory controls
• Role-based access control (RBAC)

RBAC stands for Role-Based Access Control. It manages permissions and access levels for Admin and Developer users.

Clawix currently supports:

• Admin
• Developer

Admin users can:

• Configure AI providers
• Manage communication channels
• Create and manage users
• Configure Telegram integration
• Manage organization settings and governance policies

Developer users can:

• Create and manage agents
• Create sub-agents
• Execute workflows
• Use AI communication channels

Developers cannot manage system-wide settings or create Admin accounts.

Yes. Existing Admin users can create additional Admin accounts if required.

Clawix is designed with governance principles aligned to enterprise AI governance practices, including auditability, accountability, operational oversight, and secure AI execution.

Clawix supports:

• Private memory
• Group memory
• Organization-wide memory

This enables AI agents to retain contextual knowledge across workflows and teams.

Persistent memory allows AI agents to maintain long-term context, improve continuity, and reduce repetitive prompting.

Persistent workspaces are user-specific directories that survive container restarts and store files, outputs, and project artifacts.

Skills are reusable capabilities that AI agents can execute, including:

• Web search
• File operations
• API calls
• Automation workflows

Yes. Developers can create custom skills dynamically or generate them using the built-in skill-creator agent.

Basic programming knowledge is helpful, but Clawix also supports natural-language-assisted skill generation.

Core technologies include:

• NestJS
• Next.js
• TypeScript
• PostgreSQL
• Redis
• Prisma ORM
• Docker

Yes. Clawix provides a modern TypeScript architecture, modular services, Docker-based environments, and production-grade engineering practices suitable for learning AI infrastructure development.

Contributors can:

• Submit pull requests
• Improve documentation
• Create skills
• Develop integrations
• Report bugs
• Participate in governance improvements

Yes. Clawix follows:

• TypeScript strict mode
• Conventional commits
• Automated testing
• ESLint and Prettier standards

Installation failures commonly happen when old Docker volumes still exist while new environment credentials are generated, causing mismatched database credentials.

Remove all containers, volumes, and old environment files before reinstalling:

docker compose -f docker-compose.prod.yml down --volumes --rmi all --remove-orphans

Then rerun:

pnpm run install:clawix

This usually indicates:

• Database initialization problems
• Missing POSTGRES_PASSWORD
• Old Docker volumes conflicting with new environment variables

Check the organization timezone settings inside Clawix. The server timezone and configured application timezone may differ.

No. Users can log in, log out, and switch accounts without restarting Docker services.

Planned roadmap items include:

• WhatsApp integration
• Slack integration
• Skill marketplace
• Advanced token analytics
• Multi-region deployment
• Additional AI provider integrations