openclaw
Apache 2.0
A universal LLM gateway with OpenAI, Anthropic, OpenRouter, local models and Telegram/Slack/Discord channel support.
Team and principles
AI-agent infrastructure
under your control
We build infrastructure that makes it easy to launch and operate your own AI agents: openclaw, hermes, paperclip. On your VDS. No vendor lock-in, no surprise blocks, no subscription fee for the platform itself.
Mission: Give control over AI infrastructure back to the user. Anyone comfortable with SSH should be able to bring up an agent in 5 minutes — and keep full root, audit and the right to wipe everything in a single click.
Open-source cores
hermes
MIT
A workflow orchestrator: declarative pipelines, retry policies, schedules and observability.
paperclip
Apache 2.0 (fork)
A multi-agent coordinator: roles, permissions, knowledge sharing between agents and a tamper-evident audit log.
Frequently asked questions
How is myaigent.store different from a managed cloud such as AWS Bedrock or Google Vertex?
myaigent.store installs on your own VDS. Full root access, no 'policy' lock-outs, no outages due to someone else's incident. Sources are open (Apache 2.0).
Which VDS providers are supported?
Any provider with Debian 13 or Ubuntu 24.04 LTS and root SSH access. The playbook is tested on Timeweb, Aeza, Selectel and Yandex Cloud.
Do I need my own OpenAI API key?
Not required — by default we ship a shared key with built-in rate limits. To remove limits, attach your own OpenAI / Anthropic / Ollama key — billing goes directly to the provider, no markup from us.