API ENTERPRISE

Reference API

L'API AISupportBot vous permet d'envoyer des questions a votre assistant configure et de recevoir des reponses IA basees sur votre base de connaissances. Ideal pour integration dans votre bot Discord, application web ou backend.

⚡

RAG-Powered

Answers grounded in your uploaded documents. No hallucinations.

👤

User Context

Pass custom JSON data to personalize responses per-user.

📊

Usage Tracked

Each call counts against your plan's monthly response quota.

Base URL

https://aisupportbot.io/api/chat

Authentication

All Enterprise API requests must include your API key in the request header.

Header

X-API-Key:ask_your_api_key_here

🔑

Where to find your key

Go to your Dashboard → Server → Settings. The "API Key" section is visible on Enterprise and Unlimited plans. Click Generate key to create one.

⚠️

Keep it secret

Your API key grants full access to your assistant. Never expose it in client-side code or public repositories.

Chat Endpoint

Send a message to your assistant and receive an AI-generated response.

POST /api/chat

Request Body (JSON)

Field	Type	Required	Description
assistantId	string	required	Your assistant's ID (found in the dashboard URL).
message	string	required	The user's message / question.
history	array	optional	Previous messages for context. Array of `{role, content}`. Max 15 messages kept.
userId	string	optional	Your user's identifier (for logging).
channelId	string	optional	Discord channel ID if the question came from a specific channel.
userContext	object	optional	JSON object with user-specific data injected into the AI prompt. Max 4KB. More details →

Minimal Request

curl

curl -X POST https://aisupportbot.io/api/chat \
  -H "X-API-Key: ask_your_api_key" \
  -H "Content-Type: application/json" \
  -d '{
    "assistantId": "clz1abc...",
    "message": "How do I reset my password?"
  }'

User Context

The userContext field lets you pass any JSON object about the asking user. The AI will use this data to personalize its response — addressing the user by name, referencing their account state, plan, stats, etc.

💡 Perfect for platforms with user data

If you're integrating into a game, SaaS, or community platform, pass relevant data like balance, tier, or recent activity. The AI will reference it naturally when it's relevant to the question.

Example — Bot Hosting Platform

JSON

{
  "assistantId": "clz1abc...",
  "message": "Why is my bot offline?",
  "userId": "discord_123456",
  "userContext": {
    "username": "NikolaB",
    "plan": "pro",
    "servers": [
      { "name": "MyBot", "status": "offline", "uptime": "98.2%" }
    ],
    "coins": 450,
    "memberSince": "2024-03-15"
  }
}

The AI will respond to "Why is my bot offline?" with something like:

"Hi NikolaB! I can see that MyBot (Pro plan) is currently offline. Common causes include…"

Constraints

✓ Any valid JSON object is accepted
✓ Nested objects and arrays are supported
✗ Max size: 4 KB — requests with larger payloads will be rejected
⚠ The data is read-only context — the AI cannot modify it

Response Schema

All successful responses return 200 OK with a JSON body.

Response — 200 OK

{
  "response": "The reset link is sent to your registered email…",
  "requiresEscalation": false,
  "usage": {
    "prompt_tokens": 312,
    "completion_tokens": 48,
    "total_tokens": 360
  }
}

Field	Type	Description
response	string	The AI-generated answer. May be an empty string if escalation was triggered.
requiresEscalation	boolean	`true` when the AI isn't confident enough to answer. You should route the user to a human agent.
usage	object	Token counts for the request (prompt + completion). Useful for cost monitoring.

Error Codes

All errors return a JSON body with an error field and optional details.

400

Bad Request

Missing required fields (assistantId, message) or userContext exceeds 4KB.

401

Unauthorized

API key is missing, malformed (must start with ask_), or has been deleted.

403

Forbidden

The API key does not belong to the requested assistantId.

404

Not Found

The given assistantId doesn't exist or the assistant is not active.

429

Plan Limit Reached

Your assistant has exhausted its monthly response quota. Upgrade or wait for reset.

500

Internal Server Error

Unexpected server error. Retry after a few seconds.

Error response example

{
  "error": "Plan limit reached",
  "details": "Upgrade to get more responses. (1000/1000)"
}

Examples

Full working examples for common use cases.

JS JavaScript / TypeScript

With conversation history + userContext

const askBot = async (message, user, history = []) => {
  const res = await fetch('https://aisupportbot.io/api/chat', {
    method: 'POST',
    headers: {
      'Content-Type': 'application/json',
      'X-API-Key': 'ask_your_key_here',
    },
    body: JSON.stringify({
      assistantId: 'clz1abc...',
      message,
      history,
      userId: user.id,
      userContext: {
        username: user.name,
        plan: user.plan,
        coins: user.coins,
        servers: user.servers,
      },
    }),
  });

  const data = await res.json();

  if (data.requiresEscalation) {
    return { reply: "I'll connect you with a human agent.", escalate: true };
  }

  return { reply: data.response, escalate: false };
};

Py Python

Simple integration

import requests

API_KEY = "ask_your_key_here"
ASSISTANT_ID = "clz1abc..."

def ask_bot(message, user_context=None):
    response = requests.post(
        "https://aisupportbot.io/api/chat",
        headers={
            "X-API-Key": API_KEY,
            "Content-Type": "application/json"
        },
        json={
            "assistantId": ASSISTANT_ID,
            "message": message,
            "userContext": user_context,
        }
    )
    data = response.json()

    if data.get("requiresEscalation"):
        return "Escalating to a human agent."

    return data["response"]

Handling Escalation

When requiresEscalation is true, the AI couldn't find a confident answer. You should redirect the user to a human support channel rather than showing an empty or incorrect response.

✓ Open a ticket in your Discord server

✓ Redirect to a #support channel

✓ Send a DM to on-call staff

✗ Don't show the empty response string to users

Questions about the API? Join our Discord support server.