AI Features
Ainecto has three built-in AI capabilities. They use a small monthly credit budget — see Plans for limits.
Don't confuse these with MCP integration, which is the outbound protocol that lets external AI coding tools call into Ainecto. The features on this page are the inbound AI assistant inside the Ainecto app itself.
Natural Language → ERD
Describe a database in plain English (or Korean, etc.) and Ainecto generates the schema for you.
How to use:
- Open a diagram (or create a new empty one)
- Open the AI panel
- Type something like:
A SaaS app with users, projects, tasks. Tasks belong to a project and have an assignee. Add a tags many-to-many on tasks.
- Ainecto returns a complete schema; review and accept it
The result is added to the canvas as tables, columns, and refs. A new version snapshot is created automatically with the description "Converted via AI" so you can roll back if needed.
You can also pass an existing schema and ask for refinements:
Add a soft-delete column to all tables, and add an audit_log table tracking changes.
The conversion uses 1 AI credit per call.
AI Chat
The AI Chat panel is a conversational interface that knows about your current diagram. Use it to:
- Ask questions about your schema ("which tables reference users?")
- Request changes ("add an
is_archivedcolumn to projects with a default of false") - Get migration advice
- Generate test data ideas
The chat history is preserved per diagram. Each message uses a small number of credits.
Title Suggestion
When you create a diagram and the schema is non-trivial, Ainecto can suggest a meaningful title for it based on the schema content. Look for the Suggest title button in the diagram settings.
AI Credits
AI features run on a credit budget:
| Plan | Monthly Credits |
|---|---|
| Free | None (AI disabled) |
| Pro | Included monthly allowance |
| Enterprise | Unlimited (or large allowance, custom) |
Credits reset monthly. You can also purchase top-up credit packs for one-off bursts. Check your current balance in account settings.
If you hit the credit limit or rate limit, AI calls return an error and the operation is rejected. The diagram is unchanged until the next successful call.
How is this different from MCP?
- AI Features (this page) — Built into Ainecto. You talk to the AI inside Ainecto, and Ainecto calls the LLM provider on your behalf.
- MCP Integration (overview) — Your AI coding tool (Claude Code, Cursor, etc.) calls Ainecto over MCP to read and modify the schema.
The two complement each other:
- Use AI Features when you're designing the schema in Ainecto itself
- Use MCP when your coding assistant needs to read or change the schema while writing code