Skip to content

README.md

Latest commit

 

History

History
75 lines (56 loc) · 1.74 KB

README.md

File metadata and controls

75 lines (56 loc) · 1.74 KB

Flowchart API

/api/chat

Vercel Serverless Function that proxies AI requests to Azure OpenAI (or compatible endpoint).

Environment Variables

Server-side environment variables (NOT VITE_*):

  • AZURE_DEPLOYMENT_NAME - Azure OpenAI deployment name
  • AZURE_RESOURCE_NAME - Azure OpenAI resource name
  • AZURE_API_KEY - Azure OpenAI API key

Request Format

POST /api/chat

{
  "messages": [
    { "role": "user", "content": "Create a login flow" }
  ],
  "flowContext": {
    "nodes": [...],  // Current flowchart nodes
    "edges": [...]   // Current flowchart edges
  }
}

Response Format

{
  "message": "```json\n{...}\n```",
  "role": "assistant"
}

The message field contains a JSON code block with the flowchart proposal.

Flowchart Generation Skill

The API uses a skill-based prompt engineering approach defined in flowchart-generation-skill.md.

This skill file:

  • Defines the exact JSON schema the AI must use
  • Provides multiple correct/wrong examples
  • Enforces strict flowchart-only output
  • Is loaded at runtime and sent as the system prompt

To modify AI behavior: Edit flowchart-generation-skill.md, not chat.ts.

Architecture

User Input → Frontend (AIChat.tsx)
    ↓
    → POST /api/chat (with flowContext)
    ↓
    → Loads flowchart-generation-skill.md
    ↓
    → Azure OpenAI API (with skill as system prompt)
    ↓
    → Returns flowchart JSON
    ↓
    → Frontend parses JSON → Shows preview → Inserts on confirm

Error Handling

  • 400 - Invalid request (missing messages)
  • 405 - Method not allowed (non-POST)
  • 500 - Server configuration error or Azure OpenAI API error
  • 200 - Success (includes assistant message)