Dino-GSP Open Platform

English

简体中文

English

简体中文

Appearance

Embedded Editor AI Drawing, API, and Usage Reporting Are Now Available ​

  • Type: New capability
  • Published: 2026-06-29
  • Effective date: 2026-06-29
  • Audience: Developers using SDK editor mode, embedded AI drawing, or Open Platform APIs
  • Action required: To enable this capability, upgrade the SDK, integrate the embedded-mode AI drawing API, and implement host-side request forwarding and quota checks

What Changed ​

This launch includes three parts:

  1. SDK 2.8.0 supports AI drawing in the embedded editor: editor mode can enable the AI Chat panel. When users start an AI drawing request, the host page listens for the SDK aiRequest event and passes the returned stream back to the editor with editor.ai.consumeStream().
  2. The embedded-mode AI drawing API is available: the host backend can forward the editor-generated AI request context to the Dino-GSP backend as-is. Dino-GSP handles model calls, geometry instruction generation, and streaming results. The browser does not need to, and should not, hold server-side API tokens.
  3. Console usage reporting is available: the Open Platform console now shows application-level AI call usage and point consumption, making quota management, billing reconciliation, and debugging easier.

The overall flow is: the user enters a request in the embedded editor, the SDK emits aiRequest, the host backend performs user auth and quota checks before calling the Dino-GSP AI drawing API, and the frontend passes the returned stream back to the SDK so the editor can update conversation and drawing state.

Billing Flow ​

Embedded editor AI drawing is bridged by the SDK. Actual billing is based on AI token usage produced by the Dino-GSP backend.

LayerNotes
Charged accountThe application account that calls Dino-GSP Open Platform AI capabilities
Charge timingAny request that produces AI token usage is charged by actual consumption; failure, cancellation, or disconnection does not automatically mean no charge
Pricing basisCalculated by actual token usage × model price
Usage visibilityCall records and point consumption are available in the Open Platform console application details
Host quotaThe host system may still enforce its own user quota, plans, usage count, or risk rules on the host backend

Use Cases ​

  • Embed the Dino-GSP editor in your own product and let users generate or modify geometry through natural language
  • Build AI tutoring, problem explanation, lesson authoring, and courseware editing workflows with geometry drawing
  • Keep authentication, billing, logging, and risk controls within the host system boundary

How to Start ​

  1. Upgrade to @dajiaoai/algeo-sdk@2.8.0 or later.
  2. Create the embedded editor with createEditor(...) and provide a valid auth.appId.
  3. Configure your application in the Open Platform console and prepare a server-side API token.
  4. Enable or keep the default ui.aiChatPanel setting.
  5. Listen for aiRequest and forward payload as-is to your host backend.
  6. After user auth and quota checks, the host backend calls the Dino-GSP embedded-mode AI drawing API.
  7. Consume the returned stream on the frontend with editor.ai.consumeStream().
  8. Review application usage and point consumption in the console after launch.