Skip to main content

Hyper AI

Hyper AI is a co-developer embedded directly into the platform. It understands your schemas, permissions, and domains to safely act on them when you ask.

Open the Hyper AI panel from the Studio sidebar.

Feature support and roadmap​

Hyper AI supports the following areas of the platform:

  • Collections & Fields
  • MCP Tools & Agents
  • Dashboards & Widgets
  • Areas & Navigation
  • ACL & Permissions
  • Menus & Routing
  • Collection Configuration
  • Automations

Panel controls​

Use the Hyper AI panel header to manage chat sessions and layout.

Header actions​

IconAction
New Chat — Opens a fresh chat session
Expand — Expands the AI panel to full screen
Collapse — Minimizes the AI panel
Info — Shows information about the current AI model
  • Chat — The main chat panel for asking questions and describing tasks.
  • History — View and search your previous chat sessions.

Context Mentions (@)​

Add explicit context by typing @ and mentioning specific modules directly in your prompt.

  1. Type @ in the chat input to open the Mention dropdown.
  2. Browse categories such as Collection, Role, Area, Automation, Settings, and MCP.
  3. Press Enter to drill into a category and select a specific item.
  4. Press Shift + Enter to select the category directly without drilling in.

Selected mentions appear as chips in your chat input, so Hyper AI focuses on exactly what you specified.

Build with Hyper AI​

Hyper AI is built on three core layers that give it system awareness and reasoning.

1. Knowledge layer​

Gives Hyper AI full system awareness. It continuously understands:

  • Your schemas and relationships
  • Areas, dashboards, and navigation
  • Permissions, roles, and ACLs
  • Domain patterns inferred from usage

2. Intelligence and reasoning layer​

Interprets your intent and reasons about impact, handling architectural trade-offs, performance, and compliance. Hyper AI selects the appropriate model based on task complexity.

3. Action and execution layer​

Safely applies changes to your system. Every action follows a strict lifecycle:

1

Suggest

Hyper AI proposes changes based on your natural language intent.
2

Preview

Review a visual or structured diff of the proposed modifications.
3

Validate

The system checks the proposal against schemas and permissions.
4

Apply

Changes are committed to your system after your approval.

Industry-aware by design​

Hyper AI understands the industry your system operates in — including domain workflows, operational rules, and regulatory requirements. It recommends solutions that account for industry-specific constraints, not just technical correctness.

Prompts and responses​

Use natural language to describe your goals:

  • "Build a dashboard for active orders."
  • "Add audit fields to all collections."
  • "Create an admin area with restricted access."
  • "Optimize this schema for 10M records."
  • "Add an MCP tool to list all pending orders."
  • "Set up a welcome email automation when a new user is created."

Response types​

Hyper AI responds with one of the following outcomes:

ResponseDescription
ExplanationGuidance and reasoning about your request
PreviewVisual or structured diff of proposed changes
PatchMinimal, mergeable change applied to your system
ErrorSafe rejection with an explanation of what went wrong

Getting started​

  1. Open the Hyper AI panel from the Studio sidebar.
  2. Type a task in natural language — for example, "Add a status field to the Orders collection with values pending, active, and completed."
  3. Review the Preview of proposed changes.
  4. Click Apply to commit the changes, or ask follow-up questions to refine.

Tips for better results​

  • Be specific about the collection or area — Instead of "add a field," say "add a priority field to the Tasks collection."
  • Use @ mentions — Type @ to pin a specific collection, role, or area as context before describing your task.
  • Describe the goal, not the steps — "Set up role-based access so managers can edit but viewers can only read" works better than listing every permission manually.
  • Iterate — If the first result isn't right, describe what to change. Hyper AI retains context within the session.
  • Collections — Data structure and field definitions
  • Areas — UI layouts and navigation
  • ACL — Access control and permissions
  • MCP — Expose your app to AI tools like Claude Desktop
  • Deploy and Export — Export and deploy your project