01
Where it started
Clawga's infrastructure started with OpenClaw, an open-source agent runtime. It gave us a solid foundation: a gateway, tools, sessions and persona files.
We forked it early and have been building our own agentic layer on top of it ever since, shaped around how people actually use an assistant on their phone.
02
Clawga Ocean:1
That work became Clawga Ocean:1, our native flagship model. It is designed for agentic, multi-step orchestration: breaking a request into steps, choosing the right tools, and carrying the plan through to a finished result.
Because the model and the agent layer are built together, Ocean:1 unlocks things a plain chatbot cannot: bigger tasks handled step by step, reminders it schedules for you, and skills that develop as you use it.
03
What happens when you ask
- Understand: Ocean:1 reads your request and the context of the chat.
- Plan: bigger asks are broken into clear steps.
- Act: it uses tools such as research, image generation and reminders.
- Deliver: the result lands in the chat, with the steps it took.
04
Personas: SOUL.md and AGENTS.md
Every agent has two plain markdown files. SOUL.md sets its voice and personality. AGENTS.md sets the rules it follows while it works. Edit both from the Persona screen, and reset to the default any time.
05
Bring your own key
Want a specific model? Add your own OpenAI, Anthropic or Google API key in the app. Your keys are stored only on your device, and with your own key your chats go straight to that provider.
06
Your data
We never use your chats to train AI models. You can delete your chats, or your whole account, from inside the app at any time.