denisrog/my-moltbot

Overview: OpenClaw AI Assistant on Cloudflare Infrastructure

The denisrog/my-moltbot project represents an experimental implementation of OpenClaw (formerly known as Moltbot and Clawdbot), a personal AI assistant designed to run entirely within Cloudflare's serverless infrastructure. This TypeScript-based project leverages Cloudflare Workers and Sandboxes to deliver a fully managed, always-on AI assistant without the complexities of traditional self-hosting.

As a proof-of-concept deployment, this project demonstrates how modern AI assistants can operate within containerized edge computing environments. It's particularly noteworthy for developers seeking to understand how cloudflare-workers-ai deployments work in production scenarios, though the repository clearly marks this as experimental software that may break without notice.

Key Features and Capabilities

OpenClaw provides a comprehensive personal assistant experience through its gateway architecture. The system includes a web-based Control UI that serves as the primary chat interface, accessible directly through your deployed worker URL. Multi-channel support extends functionality to popular platforms including Telegram, Discord, and Slack, enabling users to interact with their AI assistant across different communication channels.

Security is addressed through device pairing functionality that requires explicit approval via direct message authentication. This ensures only authorized devices can access your personal assistant. The system maintains persistent conversations with chat history and context preservation across sessions, while an extensible agent runtime provides workspace capabilities and customizable skills.

Architecture and Infrastructure

The project utilizes Cloudflare's standard-1 container instance specification (1/2 vCPU, 4 GiB memory, 8 GB disk). Understanding the cost structure is essential: running 24/7, users should expect approximately $34.50/month including the required Workers Paid plan ($5/month). However, the implementation includes a configurable SANDBOX_SLEEP_AFTER parameter that allows the container to sleep during idle periods, potentially reducing costs to $5-6/month for containers running only 4 hours daily.

The system integrates with anthropic-claude for AI capabilities, supporting both direct Anthropic API access and Cloudflare AI Gateway's Unified Billing option. Optional R2 Storage enables data persistence across container restarts, ensuring paired devices and conversation history survive deployments.

Installation and Setup Process

Deployment requires a Cloudflare Workers Paid plan and either an Anthropic API key or Cloudflare AI Gateway credentials. The installation process follows standard Node.js patterns with npm for dependency management. Security configuration involves generating a gateway token using OpenSSL for remote access control, then setting it via Wrangler secrets management.

A critical post-deployment requirement involves configuring Cloudflare Access to protect the administrative interface at /_admin/. This multi-step process includes enabling Access on the workers.dev domain, configuring the CF_ACCESS_TEAM_DOMAIN and CF_ACCESS_AUD secrets, and setting up identity provider authentication. Without completing these steps, the Control UI remains inaccessible despite successful deployment.

Technical Stack Analysis

The project is built entirely in TypeScript, providing type safety and modern development tooling. The technology stack includes Vite for build tooling and Vitest for testing infrastructure. Configuration files reveal the use of oxc-based tools (.oxfmtrc.json, .oxlintrc.json) for fast formatting and linting. The presence of a Dockerfile suggests containerization capabilities beyond Cloudflare's native sandbox environment.

Integration with Cloudflare's ecosystem is extensive, utilizing Browser Rendering for web navigation capabilities, AI Gateway for API routing and analytics, and Access for authentication. The wrangler.jsonc configuration file manages deployment settings and bindings for these services.

Community and Development Status

As of February 2026, the repository shows minimal community engagement with zero stars, forks, and watchers. The project has no open issues and appears to be in early experimental stages. The Apache-2.0 license provides permissive terms for modification and distribution, encouraging experimentation despite the proof-of-concept status.

Documentation includes dedicated files for AGENTS.md and CONTRIBUTING.md, suggesting structured approaches to both usage and community contribution. The project includes test infrastructure with an e2e directory and comprehensive configuration files, indicating attention to quality assurance despite the experimental label.

Comparison with Alternative AI Assistant Deployments

Unlike traditional self-hosted AI assistants requiring dedicated servers or virtual machines, this Cloudflare-based approach offers several advantages. The serverless architecture eliminates server maintenance overhead, while edge deployment provides global availability and reduced latency. Cost predictability improves through Cloudflare's transparent pricing model, though 24/7 operation costs exceed typical $5-10/month VPS hosting.

Compared to fully managed AI assistant services, this solution provides greater customization and control while maintaining infrastructure simplicity. The multi-channel support and device pairing features differentiate it from simple chatbot-frameworks implementations, offering enterprise-grade functionality in a personal assistant package.

For developers evaluating AI deployment strategies, this project demonstrates the viability of edge computing for conversational AI while highlighting the importance of cost management through intelligent sleep configurations. The experimental nature makes it ideal for learning and prototyping rather than production-critical applications.