Posterboi004/donna-moltbot-sandbox

OpenClaw on Cloudflare Workers: A Comprehensive Review

The donna-moltbot-sandbox project represents an experimental implementation of OpenClaw (formerly known as Moltbot and Clawdbot) running entirely within Cloudflare's serverless infrastructure. This TypeScript-based project demonstrates that a fully-featured personal AI assistant can operate in a Cloudflare Sandbox environment, offering a managed, always-on deployment without traditional server maintenance.

Project Overview and Purpose

OpenClaw is a personal AI assistant built with a gateway architecture that enables connections to multiple chat platforms simultaneously. The donna-moltbot-sandbox packages this assistant to run as a Cloudflare Worker with Sandbox container support, leveraging Cloudflare's edge infrastructure for global availability and low latency. Released under the Apache-2.0 license, this proof-of-concept project is explicitly marked as experimental and not officially supported.

The project integrates with Anthropic's Claude API for natural language processing and includes features like a web-based Control UI, multi-channel support for Telegram, Discord, and Slack, secure device pairing with explicit approval requirements, persistent conversation history, and an extensible agent runtime with workspace and skills capabilities. rogerli007/my-openclaw-page

Features and Capabilities

Gateway Architecture

The core strength of this implementation lies in its gateway design, which serves as a central hub for managing multiple chat platform connections. Users can interact with their AI assistant across different platforms while maintaining conversation context and history. The Control UI provides a unified web interface for direct interaction at the gateway level.

Security and Authentication

Security is implemented through multiple layers. The project requires Cloudflare Access configuration to protect the admin UI, ensuring only authorized users can manage device pairings. Device pairing requires explicit approval through an admin interface located at the /_admin/ endpoint, preventing unauthorized access to the assistant. Gateway token authentication secures remote access to the Control UI. cloudflare-access-setup

Optional Persistence

While the sandbox container is ephemeral by default, the project supports optional R2 storage integration for persisting paired devices and conversation history across container restarts. This is highly recommended for production use to maintain continuity of service.

Technical Stack Analysis

The project is built entirely in TypeScript, utilizing modern tooling for development and deployment. Key technologies include:

• Cloudflare Workers: Serverless execution environment running at the edge
• Cloudflare Sandbox: Container runtime with standard-1 instance (1/2 vCPU, 4 GiB memory, 8 GB disk)
• Cloudflare Access: Zero Trust authentication for admin UI protection
• Cloudflare Browser Rendering: For browser navigation capabilities
• Cloudflare AI Gateway: Optional unified billing and API routing for AI services
• Cloudflare R2: Optional object storage for persistence
• Vite: Build tooling and development server
• Vitest: Testing framework

The project includes configuration files for oxc (fast JavaScript tooling), Docker containerization, and Wrangler for Cloudflare Workers deployment. cloudflare-workers

Installation and Setup Process

Setting up the donna-moltbot-sandbox requires a Workers Paid plan ($5/month) and an Anthropic API key. The deployment process involves installing dependencies via npm, configuring secrets through Wrangler CLI (including ANTHROPIC_API_KEY and MOLTBOT_GATEWAY_TOKEN), and deploying with npm run deploy.

Post-deployment configuration is critical and includes enabling Cloudflare Access on the workers.dev domain, setting CF_ACCESS_TEAM_DOMAIN and CF_ACCESS_AUD secrets for JWT validation, and configuring access policies in the Zero Trust Dashboard. The first container startup can take 1-2 minutes, and device pairing through the admin UI is mandatory before the Control UI becomes functional.

Cost Considerations

Running this project 24/7 costs approximately $34.50/month, including the $5 Workers Paid plan fee. The breakdown includes ~$26/month for 4 GiB memory provisioning, ~$2/month for CPU at 10% utilization, and ~$1.50/month for 8 GB disk storage. These costs are based on Cloudflare's pay-as-you-go pricing beyond free tier allocations.

Cost optimization is possible through the SANDBOX_SLEEP_AFTER configuration parameter, which puts the container to sleep during idle periods. A container running only 4 hours daily would cost approximately $5-6/month in compute resources plus the base plan fee, making it economically viable for personal use. cloudflare-workers-cost-optimization

Community and Development Status

As of February 2026, this repository shows minimal community engagement with 0 stars, 0 forks, and 0 open issues. The project was last updated on February 7, 2026, and contains only 1 commit, indicating it's in very early stages. The lack of GitHub topics makes discoverability challenging. However, this early-stage status also presents opportunities for contributors interested in serverless AI assistant architectures.

Comparison with Alternatives

Traditional self-hosted AI assistants require dedicated servers or VPS instances, typically costing $10-50/month with significant maintenance overhead. The donna-moltbot-sandbox offers comparable functionality at competitive pricing while eliminating server management tasks. Compared to cloud-hosted AI platforms like those from major providers, this solution provides greater control and customization while leveraging Cloudflare's global edge network for low-latency responses. ai-assistant-alternatives

The trade-off is the experimental nature and limited support—production users should be prepared to troubleshoot issues independently. For developers comfortable with TypeScript and Cloudflare's ecosystem, this project offers an intriguing middle ground between fully managed services and self-hosted solutions.

Conclusion

The donna-moltbot-sandbox successfully demonstrates that sophisticated AI assistants can run entirely on serverless infrastructure. While explicitly experimental, the project showcases the potential of edge computing for AI applications. The comprehensive setup process, reasonable cost structure with optimization options, and extensible architecture make it worth exploring for developers seeking alternatives to traditional deployment models. However, the lack of community activity and official support means this is best suited for technical users willing to embrace bleeding-edge experimentation.