This article is still being authored. This feature is currently under development. Content will be updated as the feature is finalized.

OpenAI Model Endpoint Facades

Learn about FoundationaLLM's OpenAI-compatible endpoint facades for simplified model integration.

Overview

TODO: This feature is under active development. Documentation will be completed when the feature is released.

OpenAI Model Endpoint Facades provide an OpenAI-compatible API layer in front of FoundationaLLM's agent capabilities. This allows applications built for the OpenAI API to integrate with FoundationaLLM with minimal code changes.

Purpose and Use Cases

Why Use Endpoint Facades?

Benefit	Description
Compatibility	Use existing OpenAI SDK code
Migration Path	Gradually move from OpenAI to FoundationaLLM
Familiar Interface	Standard chat completion format
Agent Benefits	Access FoundationaLLM's agent capabilities

Ideal Scenarios

Migrating existing OpenAI-integrated applications
Using libraries that expect OpenAI-compatible endpoints
Building applications that may switch between providers
Teams familiar with OpenAI API patterns

Feature Status

Component	Status
Chat Completions endpoint	Under Development
Embeddings endpoint	Under Development
Streaming responses	Under Development
Function calling	Under Development

Configuration

TODO: Document configuration steps once the feature is finalized, including:

Enabling endpoint facades in deployment

Configuring facade endpoints

Mapping facades to agents

Authentication configuration

Prerequisites

FoundationaLLM deployment with endpoint facades enabled
Appropriate API credentials configured
Agent(s) to expose via the facade

Setup Steps

TODO: Provide step-by-step setup instructions

API Reference

Chat Completions Endpoint

TODO: Document the endpoint specification

Expected Endpoint:

POST /v1/chat/completions

Expected Request Format:

{
  "model": "agent-name",
  "messages": [
    {"role": "system", "content": "You are a helpful assistant."},
    {"role": "user", "content": "Hello!"}
  ],
  "temperature": 0.7,
  "max_tokens": 1000
}

Expected Response Format:

{
  "id": "chatcmpl-xxx",
  "object": "chat.completion",
  "created": 1234567890,
  "model": "agent-name",
  "choices": [
    {
      "index": 0,
      "message": {
        "role": "assistant",
        "content": "Hello! How can I help you today?"
      },
      "finish_reason": "stop"
    }
  ],
  "usage": {
    "prompt_tokens": 10,
    "completion_tokens": 15,
    "total_tokens": 25
  }
}

Embeddings Endpoint

TODO: Document embeddings endpoint if included

Usage Examples

Python with OpenAI SDK

TODO: Verify and complete example

from openai import OpenAI

# Configure client to use FoundationaLLM facade
client = OpenAI(
    base_url="https://{fllm-api}/v1",
    api_key="your-api-key"  # TODO: Document key format
)

# Use standard OpenAI SDK methods
response = client.chat.completions.create(
    model="your-agent-name",  # Maps to FLLM agent
    messages=[
        {"role": "user", "content": "What can you help me with?"}
    ]
)

print(response.choices[0].message.content)

curl Example

TODO: Verify and complete example

curl -X POST "https://{fllm-api}/v1/chat/completions" \
  -H "Authorization: Bearer your-api-key" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "your-agent-name",
    "messages": [
      {"role": "user", "content": "Hello!"}
    ]
  }'

Mapping to FoundationaLLM Concepts

OpenAI Concept	FoundationaLLM Equivalent
`model` parameter	Agent name
`messages` array	Conversation context
`temperature`	Model parameter override
`max_tokens`	`max_new_tokens` parameter
API Key	Agent Access Token or Bearer token

Limitations

TODO: Document known limitations once feature is finalized

Expected Limitations:

Not all OpenAI parameters may be supported
Some agent features may not map to OpenAI concepts
Streaming behavior may differ
Rate limits follow FoundationaLLM quotas

Migration Guide

From OpenAI to FoundationaLLM Facades

TODO: Provide migration steps

Configure FoundationaLLM endpoint facade
Update base URL in your application
Map model names to agent names
Test and verify functionality

Considerations

Review agent configurations match expected behavior
Test streaming if used
Verify token counting differences
Update error handling for FLLM-specific errors

Troubleshooting

Common Issues

TODO: Document common issues and solutions

Issue	Possible Cause	Solution
404 Not Found	Facade not enabled	Enable in configuration
Authentication Error	Invalid credentials	Check API key configuration
Model Not Found	Agent doesn't exist	Verify agent name

Table of Contents