Your self-hosted AI assistant
Your assistant. Your services.
Your personal Hub.
Give your agents the context to help: conversations, people, tasks and calendars. Build your way of organizing life on Gravify, with compatible services and assistants you already use.
Connect your agent to WhatsApp →Build the workflow that works for you.
Bring a conversation into Assistant chat, ask for help organizing a project, or connect an external agent to authorized information on your Hub. Automations you configure in your agent can use Gravify as their context and workspace.
Start with a single request or build a workflow you can reuse. What you can read or change in external services depends on the supported connections and their permissions.
- From a group chat to a project. Gather decisions and open questions, prepare a report and organize the tasks that emerge.
- From commitments to your week. Bring tasks and appointments together and ask for a proposal to make time for what you want to do.
- From context to a reply. Have the agent prepare a draft based on the conversation and your writing style. Review it before using it.
These are starting points. You choose the goal, the context and the work you give your agent.
Prepared work, in one place to review.
Drafts stay connected to their conversations. Open the context, edit the text or discard the proposal. For a direct chat, you can open the reply in WhatsApp; sending remains your choice.

See what will change in your calendar.
Calendar proposals show the change for you to review before applying it. Check the placement, then choose whether to apply or discard the proposal.

You choose the context.
Allow access to the people and chats useful for your agent’s work, exclude those you want to keep out and control accessible calendars. The rules apply to the Assistant and connected agents.
See how permissions work →Local AI, even offline.
Use the built-in Assistant with a compatible local model server. With the model downloaded and the server running, it can work on locally stored content offline. Install models and the server separately. External agents and cloud models remain optional; context they use goes to their provider.
Explore private local AI →Use the assistant you prefer.
Chat inside Gravify through a supported Codex connection or an OpenRouter model, or connect Codex, Claude Code or another compatible local MCP agent. Your computer and Personal Hub must remain running for agent access.
Your Hub keeps the private archive on hardware you control. If you choose an AI provider, the context needed for your request is shared with that provider. Its limits and costs apply.
Talk to your assistant through WhatsApp.
In the Tester channel, set up a dedicated assistant number and message it from your verified personal number. Local setup guides you through pairing and the first reply. This optional, unofficial integration requires risk acceptance and local history storage. Your Hub must stay running; the number and AI service have separate costs. It replies only to you, with no groups, other recipients or proactive messages.
This account is separate from the connector that imports your conversations. Drafts for other people remain subject to approval and manual sending.