Platform Overview
Hatcher hosts OpenClaw and Hermes agents in isolated Docker containers. You configure the agent from the dashboard, Hatcher writes the native framework config, and the runtime runs with persistent storage, logs, integrations, and lifecycle management.
Runtime Flow
- You create an agent in Chat-to-Hatch.
- Hatcher generates the framework files (
openclaw.json,config.yaml,SOUL.md,.env, integration config). - Hosted model calls route through the Hatcher LLM proxy, or BYOK keys are injected directly into the container.
- The agent runs in an isolated Docker network with CPU, RAM, workspace, and restart limits.
- Hatcher tracks status, logs, files, billing, AI Credit usage, and integrations.
Frameworks
| OpenClaw | Hermes | |
|---|---|---|
| Best for | Automation, multi-channel routing, ClawHub skills | Conversational agents, persistent memory, tool-heavy workflows |
| Config | openclaw.json + SOUL.md | config.yaml + .env + SOUL.md |
| Runtime port | 18789 | 8642 |
| Skills/tools | ClawHub and built-in tools | Built-in toolsets and skills |
| Memory | Markdown memory files and vector search | MEMORY.md, USER.md, session memory |
| Integrations | Telegram, Discord, WhatsApp, Slack, X/Twitter, webhooks, and more | Telegram, Discord, WhatsApp, Slack, X/Twitter, webhooks, and more |
Resource Limits
| Plan | CPU | RAM | Workspace | Auto-sleep |
|---|---|---|---|---|
| Free | 1 CPU | 1 GB | 2 GB | 12h idle |
| Starter | 1 CPU | 1.5 GB | 10 GB | Always-on |
| Pro | 1.5 CPU | 4 GB | 25 GB | Always-on |
| Business | 4 CPU | 6 GB | 50 GB | Always-on |
| Founding Member | 4 CPU | 6 GB | 40 GB | Always-on |
Workspace values are quotas. Hatcher does not reserve the full quota upfront for every agent; storage is enforced as the workspace grows.
Hosted Models And BYOK
Hosted usage routes through UsePod first with OpenRouter fallback and spends AI Credits. The agent config UI uses the same provider/model selector for OpenClaw and Hermes.
Hosted provider coverage includes DeepSeek, OpenAI, Anthropic, Google, Qwen, Moonshot AI, xAI, Mistral, Z.ai, NVIDIA, Meta, and OpenRouter fallback.
BYOK supports OpenAI, Anthropic, Google, Groq, xAI, OpenRouter, and Venice AI. BYOK usage goes to your provider account and does not spend Hatcher AI Credits.
Included Platform Features
Included on every tier:
- File Manager
- Full Logs
- Skills and plugins
- Platform integrations
- Knowledge files
- Web search capability through hosted AI Credits or BYOK provider keys
- Developer workflows: terminal sessions, GitHub runtime access, A2A, and My Agents orchestration view
Business is the highest monthly capacity tier. Founding Member is a lifetime community plan with Founding Member Discord role and Telegram tag.
Security Gates
- Containers run without Docker socket access and without host filesystem access.
- Agent workspaces are isolated per agent.
- Hosted AI usage is authorized through Hatcher’s LLM proxy and account AI Credit balance.
- BYOK keys are encrypted at rest and injected only into the owning agent runtime.
- Free agents auto-sleep after idle time; paid agents remain always-on.