Chat API

The Prometheux documentation includes an AI-powered chat API that provides Vadalog code assistance and answers questions about the platform. This API is available at https://chat-docs.prometheux.ai/api/docsChat and can be integrated into your applications.

---

Overview

The Chat API provides:

• Vadalog code generation tailored to your specific question
• Syntax assistance and debugging help
• Documentation search integration via Algolia
• Context-aware responses based on Prometheux documentation
• Grammar-compliant code following Vadalog best practices
• Intelligent examples - simple questions get focused snippets, complex questions get full programs

---

Endpoints

/api/docsChat (Streaming)

Purpose: Optimized for chat interfaces and streaming responses.

• URL: https://chat-docs.prometheux.ai/api/docsChat
• Method: POST
• Content-Type: application/json
• Response: Streaming text (AI SDK format)

Request Format

{
  "messages": [
    {
      "role": "user",
      "content": "Show me a PostgreSQL connection example"
    }
  ]
}

Response Format

Streaming text response in AI SDK format:

0:"Below"
0:" is"
0:" a"
0:" complete"
0:" Vadalog"
0:" example..."

Example Usage

JavaScript/TypeScript:

const response = await fetch('https://chat-docs.prometheux.ai/api/docsChat', {
  method: 'POST',
  headers: {
    'Content-Type': 'application/json',
  },
  body: JSON.stringify({
    messages: [
      { role: 'user', content: 'How do I connect to Neo4j?' }
    ]
  })
});

// Handle streaming response
const reader = response.body.getReader();
const decoder = new TextDecoder();
let result = '';

while (true) {
  const { done, value } = await reader.read();
  if (done) break;
  
  const chunk = decoder.decode(value, { stream: true });
  // Parse AI SDK format: 0:"text"
  const lines = chunk.split('\n');
  for (const line of lines) {
    if (line.startsWith('0:"') && line.endsWith('"')) {
      result += line.slice(3, -1); // Extract text content
    }
  }
}

console.log(result); // Complete Vadalog code example

Python:

import requests
import re

url = "https://chat-docs.prometheux.ai/api/docsChat"
payload = {
    "messages": [
        {"role": "user", "content": "Show me aggregation examples"}
    ]
}

response = requests.post(url, json=payload, stream=True)

result = ""
for chunk in response.iter_content(chunk_size=1024, decode_unicode=True):
    if chunk:
        # Parse AI SDK format: 0:"text"
        lines = chunk.split('\n')
        for line in lines:
            if line.startswith('0:"') and line.endswith('"'):
                result += line[3:-1]  # Extract text content

print(result)  # Complete Vadalog code

cURL:

curl -X POST https://chat-docs.prometheux.ai/api/docsChat \
  -H "Content-Type: application/json" \
  -d '{
    "messages": [
      {"role": "user", "content": "How do I use math functions?"}
    ]
  }'

---

/api/vadalog (Standard JSON)

Purpose: Standard REST API for programmatic integration.

• URL: https://chat-docs.prometheux.ai/api/vadalog
• Method: POST
• Content-Type: application/json
• Response: Standard JSON

Request Format

{
  "query": "Show me a PostgreSQL connection example",
  "context": "database connection",
  "include_docs": true
}

Response Format

{
  "response": "Below is a complete Vadalog example...",
  "code_examples": [
    {
      "language": "vadalog",
      "code": "@bind(\"customer\", \"postgresql host=localhost port=5432 user=myuser password=mypass\", \"mydb\", \"customer\").\n@model(\"customer\", \"['id:int', 'name:string', 'age:int']\").\n@output(\"result\").",
      "description": "PostgreSQL connection with customer data processing"
    }
  ],
  "relevant_docs": [
    {
      "title": "Connecting to Databases", 
      "url": "https://api.prometheux.ai/docs/learn/vadalog/data-sources",
      "excerpt": "PostgreSQL database connections..."
    }
  ],
  "metadata": {
    "provider": "Azure OpenAI",
    "model": "gpt-4o",
    "tokens_used": 1247,
    "search_results": 3
  }
}

Example Usage

JavaScript/TypeScript:

const response = await fetch('https://chat-docs.prometheux.ai/api/vadalog', {
  method: 'POST',
  headers: {
    'Content-Type': 'application/json',
  },
  body: JSON.stringify({
    query: 'How do I use aggregations in Vadalog?',
    include_docs: true
  })
});

const data = await response.json();
console.log(data.response);        // Text response
console.log(data.code_examples);   // Extracted code blocks
console.log(data.relevant_docs);   // Related documentation

Python:

import requests

url = "https://chat-docs.prometheux.ai/api/vadalog"
payload = {
    "query": "Show me CSV file processing",
    "context": "data integration",
    "include_docs": True
}

response = requests.post(url, json=payload)
data = response.json()

print(f"Response: {data['response']}")
print(f"Code examples: {len(data['code_examples'])}")
print(f"Related docs: {len(data['relevant_docs'])}")

cURL:

curl -X POST https://chat-docs.prometheux.ai/api/vadalog \
  -H "Content-Type: application/json" \
  -d '{
    "query": "Show me CSV file processing",
    "context": "data integration", 
    "include_docs": true
  }'

---

Features

Vadalog Code Generation

• Focused, relevant code examples matched to your question
• Proper syntax with appropriate annotations (only when needed)
• Database connection examples for data source questions
• Full data processing workflows for integration scenarios
• Simple syntax examples for concept/logic questions
• Grammar-compliant code

Documentation Integration (RAG)

• Uses Retrieval Augmented Generation (RAG) via Algolia
• Dynamically searches documentation based on your query
• Retrieves top 3 most relevant sections
• Always up-to-date - automatically includes new documentation
• No manual prompt updates needed when docs change

AI-Powered Assistance

• Powered by Azure OpenAI (gpt-4o)
• Context-aware responses
• Iterative conversation support
• Syntax error prevention

---

Response Types

Streaming Response (/api/docsChat)

• Best for: Chat interfaces, real-time applications
• Format: AI SDK compatible streaming
• Use case: Frontend chat components

JSON Response (/api/vadalog)

• Best for: Programmatic integration, APIs, automation
• Format: Standard JSON with structured data
• Use case: Backend services, integrations, analysis

---

Rate Limits

• Rate limit: 180 requests per minute
• Token limit: 30,000,000 tokens per minute
• Response limit: 1500 tokens per response
• Concurrent requests: Supported

---

Error Handling

Error Response Format

{
  "error": "AI assistant not available",
  "message": "Service temporarily unavailable",
  "status": 503,
  "timestamp": "2024-01-15T10:30:00Z"
}

Common Error Codes

Code	Description	Solution
400	Bad Request	Check request format
429	Rate Limited	Reduce request frequency
503	Service Unavailable	Retry after delay
500	Internal Error	Contact support

---

Best Practices

1. Use conversation context: Include previous messages for better responses
2. Be specific: Ask detailed questions about Vadalog syntax
3. Match question complexity:
   • For syntax questions: "How do I use the and() function?"
   • For complete programs: "Show me a complete PostgreSQL connection example"
4. Handle streaming: Implement proper streaming for real-time responses
5. Error handling: Implement retry logic for rate limits

---

Examples

Database Connections

JavaScript/TypeScript:

const response = await fetch('https://chat-docs.prometheux.ai/api/docsChat', {
  method: 'POST',
  headers: { 'Content-Type': 'application/json' },
  body: JSON.stringify({
    messages: [
      { role: 'user', content: 'How do I connect to PostgreSQL with authentication?' }
    ]
  })
});

// Parse streaming response
const reader = response.body.getReader();
const decoder = new TextDecoder();
let vadalogCode = '';

while (true) {
  const { done, value } = await reader.read();
  if (done) break;
  
  const chunk = decoder.decode(value, { stream: true });
  const lines = chunk.split('\n');
  for (const line of lines) {
    if (line.startsWith('0:"') && line.endsWith('"')) {
      vadalogCode += line.slice(3, -1);
    }
  }
}

console.log(vadalogCode); // Complete @bind example with connection parameters

Python:

import requests

url = "https://chat-docs.prometheux.ai/api/docsChat"
payload = {
    "messages": [
        {"role": "user", "content": "How do I connect to PostgreSQL with authentication?"}
    ]
}

response = requests.post(url, json=payload, stream=True)

vadalog_code = ""
for chunk in response.iter_content(chunk_size=1024, decode_unicode=True):
    if chunk:
        lines = chunk.split('\n')
        for line in lines:
            if line.startswith('0:"') and line.endswith('"'):
                vadalog_code += line[3:-1]

print(vadalog_code)  # Complete @bind example with connection parameters

cURL:

curl -X POST https://chat-docs.prometheux.ai/api/docsChat \
  -H "Content-Type: application/json" \
  -d '{
    "messages": [
      {"role": "user", "content": "How do I connect to PostgreSQL with authentication?"}
    ]
  }'

Data Processing with Aggregations

JavaScript/TypeScript:

const response = await fetch('https://chat-docs.prometheux.ai/api/vadalog', {
  method: 'POST',
  headers: { 'Content-Type': 'application/json' },
  body: JSON.stringify({
    query: 'Show me aggregation examples with grouping',
    context: 'data analysis',
    include_docs: true
  })
});

const data = await response.json();
console.log('Response:', data.response);
console.log('Code Examples:', data.code_examples);
data.code_examples.forEach(example => {
  console.log(`\n--- ${example.description} ---`);
  console.log(example.code);
});

Python:

import requests

url = "https://chat-docs.prometheux.ai/api/vadalog"
payload = {
    "query": "Show me aggregation examples with grouping",
    "context": "data analysis",
    "include_docs": True
}

response = requests.post(url, json=payload)
data = response.json()

print(f"Response: {data['response']}")
for example in data['code_examples']:
    print(f"\n--- {example['description']} ---")
    print(example['code'])

cURL:

curl -X POST https://chat-docs.prometheux.ai/api/vadalog \
  -H "Content-Type: application/json" \
  -d '{
    "query": "Show me aggregation examples with grouping",
    "context": "data analysis",
    "include_docs": true
  }' | jq '.code_examples[].code'

Advanced AI Features

JavaScript/TypeScript:

const response = await fetch('https://chat-docs.prometheux.ai/api/docsChat', {
  method: 'POST',
  headers: { 'Content-Type': 'application/json' },
  body: JSON.stringify({
    messages: [
      { role: 'user', content: 'How do I use LLM functions and embeddings in Vadalog?' }
    ]
  })
});

// Handle conversation context
const conversationHistory = [
  { role: 'user', content: 'How do I use LLM functions and embeddings in Vadalog?' }
];

// Parse response and continue conversation
const reader = response.body.getReader();
const decoder = new TextDecoder();
let assistantResponse = '';

while (true) {
  const { done, value } = await reader.read();
  if (done) break;
  
  const chunk = decoder.decode(value, { stream: true });
  const lines = chunk.split('\n');
  for (const line of lines) {
    if (line.startsWith('0:"') && line.endsWith('"')) {
      assistantResponse += line.slice(3, -1);
    }
  }
}

// Add to conversation history for follow-up questions
conversationHistory.push({ role: 'assistant', content: assistantResponse });
console.log('AI Functions Example:', assistantResponse);

Python:

import requests

url = "https://chat-docs.prometheux.ai/api/docsChat"

# Start conversation
conversation = [
    {"role": "user", "content": "How do I use LLM functions and embeddings in Vadalog?"}
]

response = requests.post(url, json={"messages": conversation}, stream=True)

assistant_response = ""
for chunk in response.iter_content(chunk_size=1024, decode_unicode=True):
    if chunk:
        lines = chunk.split('\n')
        for line in lines:
            if line.startswith('0:"') and line.endswith('"'):
                assistant_response += line[3:-1]

# Add to conversation for follow-up
conversation.append({"role": "assistant", "content": assistant_response})
print("AI Functions Example:", assistant_response)

# Follow-up question
conversation.append({"role": "user", "content": "Can you show me a similarity analysis example?"})
follow_up = requests.post(url, json={"messages": conversation}, stream=True)
# Process follow-up response...

cURL:

# Initial request
curl -X POST https://chat-docs.prometheux.ai/api/docsChat \
  -H "Content-Type: application/json" \
  -d '{
    "messages": [
      {"role": "user", "content": "How do I use LLM functions and embeddings in Vadalog?"}
    ]
  }'

# Follow-up with conversation context
curl -X POST https://chat-docs.prometheux.ai/api/docsChat \
  -H "Content-Type: application/json" \
  -d '{
    "messages": [
      {"role": "user", "content": "How do I use LLM functions and embeddings in Vadalog?"},
      {"role": "assistant", "content": "Previous response..."},
      {"role": "user", "content": "Can you show me a similarity analysis example?"}
    ]
  }'

---

Integration

Frontend Integration

Use the streaming endpoint with AI SDK components for real-time chat interfaces.

Backend Integration

Use the JSON endpoint for programmatic access, automation, and service integration.

Documentation Bot

Embed the chat functionality directly in your documentation using the provided components.

The Chat API provides comprehensive Vadalog assistance powered by AI and integrated with the complete Prometheux documentation.