Clawdi

Clawdi is a cloud-hosted AI agent environment — not an agent framework itself, but the persistent layer that sits above any agent framework. The core problem it solves is framework fragmentation: every time a developer or operator switches agent engines (OpenClaw to Hermes, Hermes to Claude Code), they lose everything — memory, API keys, skills, cron schedules, and app connections. Clawdi decouples the environment from the engine so your setup carries over regardless of which framework you run. Each workspace runs inside an Intel TDX hardware-encrypted virtual machine, giving users cryptographically verifiable privacy guarantees rather than policy promises. The agent gets a sandboxed file system and sandboxed browser. API keys are encrypted at the browser before leaving the device. Clawdi supports OpenClaw, Hermes, Claude Code, and Codex as swappable agent engines, with BYOA (bring your own agent) in development. App connections — Slack, Discord, Gmail, GitHub, Telegram, WhatsApp, and 13+ messaging platforms — are configured once in the Clawdi dashboard and surface inside agents automatically over MCP. Memory, vault, and connector tools are served through the Model Context Protocol, making Clawdi compatible with any MCP-aware agent. Automation runs via cron jobs for recurring tasks and persistent sessions for long-running workflows. The ClawHub marketplace provides community-built skills installable across all registered agents simultaneously. Pricing: Free tier includes 5,000 credits/month with one agent engine (OpenClaw or Hermes), 2 vCPU, 4GB RAM, 20GB storage, and 13+ messaging integrations. Pro is $29/month (30,000 credits, both engines, web terminal access). Max is $99/month (120,000 credits, 4 vCPU/8GB RAM). Enterprise pricing available with SSO, audit logs, and custom regions. Credits are consumed per agent action and do not roll over month to month — validate expected usage on the free tier before committing to a paid plan. No SOC 2 certification at time of listing.