aihappey-ai

A multi-provider .NET AI gateway exposing normalized endpoints for models, media, skills, agents and MCP capabilities.

No smart routing fairy. No shiny admin portal. No hidden logic maze. Just provider capabilities flattened, normalized hard and exposed through open contracts. Stateless. Boring on purpose.

Access 130k+ models and provider-native capabilities from your favorite client.

Endpoints

Request model

Requests use provider-prefixed model ids and provider-specific API keys.

Example model ids:

openai/gpt-5.4-mini
anthropic/claude-sonnet-4-5
google/gemini-2.5-pro
groq/llama-3.3-70b-versatile
openrouter/openai/gpt-5.4-mini
zai/glm-4.6

Example provider headers:

X-OpenAI-Key: <key>
X-Anthropic-Key: <key>
X-Google-Key: <key>
X-Groq-Key: <key>
X-OpenRouter-Key: <key>
X-ZAI-Key: <key>

Quick start

BASE_URL="https://ai.aihappey.net"
API_KEY="<your-provider-key>"

curl "$BASE_URL/v1/chat/completions" \
  -H "Content-Type: application/json" \
  -H "X-OpenAI-Key: $API_KEY" \
  -d '{
    "model": "openai/gpt-5.4-mini",
    "messages": [
      { "role": "user", "content": "Say hi" }
    ]
  }'

Examples

POST /api/chat

AI SDK UI compatible chat stream.

Minimal text message:

curl "$BASE_URL/api/chat" \
  -H "Content-Type: application/json" \
  -H "X-OpenAI-Key: $API_KEY" \
  -d '{
    "id": "chat-1",
    "model": "openai/gpt-5.4-mini",
    "messages": [
      {
        "id": "msg-1",
        "role": "user",
        "parts": [
          { "type": "text", "text": "Hello from AIHappey" }
        ]
      }
    ]
  }'

Tool-call capable request:

curl "$BASE_URL/api/chat" \
  -H "Content-Type: application/json" \
  -H "X-OpenAI-Key: $API_KEY" \
  -d '{
    "id": "chat-2",
    "model": "openai/gpt-5.4-mini",
    "toolChoice": "auto",
    "maxToolCalls": 1,
    "tools": [
      {
        "name": "get_weather",
        "description": "Get the current weather for a city",
        "inputSchema": {
          "type": "object",
          "properties": {
            "city": { "type": "string" }
          },
          "required": ["city"]
        }
      }
    ],
    "messages": [
      {
        "id": "msg-2",
        "role": "user",
        "parts": [
          { "type": "text", "text": "What is the weather in Amsterdam?" }
        ]
      }
    ]
  }'

POST /v1/chat/completions

OpenAI-compatible chat completions.

Non-streaming:

curl "$BASE_URL/v1/chat/completions" \
  -H "Content-Type: application/json" \
  -H "X-OpenAI-Key: $API_KEY" \
  -d '{
    "model": "openai/gpt-5.4-mini",
    "messages": [
      { "role": "user", "content": "Say hi" }
    ]
  }'

Streaming:

curl "$BASE_URL/v1/chat/completions" \
  -H "Content-Type: application/json" \
  -H "X-OpenAI-Key: $API_KEY" \
  -d '{
    "model": "openai/gpt-5.4-mini",
    "stream": true,
    "messages": [
      { "role": "user", "content": "Stream a short response" }
    ]
  }'

POST /v1/responses

OpenAI Responses-compatible endpoint.

Non-streaming:

curl "$BASE_URL/v1/responses" \
  -H "Content-Type: application/json" \
  -H "X-OpenAI-Key: $API_KEY" \
  -d '{
    "model": "openai/gpt-5.4-mini",
    "input": "List 3 creative project names"
  }'

Streaming:

curl "$BASE_URL/v1/responses" \
  -H "Content-Type: application/json" \
  -H "X-OpenAI-Key: $API_KEY" \
  -d '{
    "model": "openai/gpt-5.4-mini",
    "stream": true,
    "input": "Stream a 2-sentence summary about AIHappey"
  }'

GET /v1/models

List available models.

curl "$BASE_URL/v1/models" \
  -H "X-OpenAI-Key: $API_KEY"

GET /v1/skills

List available skills.

curl "$BASE_URL/v1/skills" \
  -H "X-OpenAI-Key: $API_KEY"

POST /api/rerank

Use a reranking-capable model.

curl "$BASE_URL/api/rerank" \
  -H "Content-Type: application/json" \
  -H "X-Cohere-Key: $API_KEY" \
  -d '{
    "model": "cohere/rerank-english-v3.0",
    "query": "best pizza in Amsterdam",
    "topN": 3,
    "documents": {
      "type": "text",
      "values": [
        "Try authentic Neapolitan pizza downtown.",
        "A cozy spot with wood-fired crusts.",
        "Grab a quick slice near the station."
      ]
    }
  }'

POST /v1/images/generations

Generate images.

curl "$BASE_URL/v1/images/generations" \
  -H "Content-Type: application/json" \
  -H "X-OpenAI-Key: $API_KEY" \
  -d '{
    "model": "openai/gpt-image-1",
    "prompt": "A minimal studio apartment in Scandinavian style",
    "size": "1024x1024",
    "n": 1
  }'

POST /v1/audio/speech

Generate speech.

curl "$BASE_URL/v1/audio/speech" \
  -H "Content-Type: application/json" \
  -H "X-OpenAI-Key: $API_KEY" \
  -d '{
    "model": "openai/tts-1",
    "voice": "alloy",
    "response_format": "mp3",
    "input": "AIHappey makes it easy to route across providers."
  }' > speech.mp3

POST /api/speech

Generate speech.

curl "$BASE_URL/api/speech" \
  -H "Content-Type: application/json" \
  -H "X-OpenAI-Key: $API_KEY" \
  -d '{
    "model": "openai/tts-1",
    "voice": "alloy",
    "outputFormat": "mp3",
    "text": "AIHappey makes it easy to route across providers."
  }' > speech.json

jq -r '.audio.base64' speech.json | base64 --decode > speech.mp3

POST /api/transcriptions

Transcribe audio.

curl "$BASE_URL/api/transcriptions" \
  -H "Content-Type: application/json" \
  -H "X-OpenAI-Key: $API_KEY" \
  -d '{
    "model": "openai/whisper-1",
    "mediaType": "audio/mpeg",
    "audio": "data:audio/mpeg;base64,<base64-audio>"
  }'

POST /v1/realtime/client_secrets

Mint realtime client secrets.

curl "$BASE_URL/v1/realtime/client_secrets" \
  -H "Content-Type: application/json" \
  -H "X-OpenAI-Key: $API_KEY" \
  -d '{
    "model": "openai/gpt-4o-realtime-preview",
    "providerOptions": {
      "openai": {
        "session": {
          "instructions": "You are a concise assistant."
        }
      }
    }
  }'

POST /api/videos

Generate videos.

curl "$BASE_URL/api/videos" \
  -H "Content-Type: application/json" \
  -H "X-OpenAI-Key: $API_KEY" \
  -d '{
    "model": "openai/sora-2",
    "prompt": "Drone shot in a minimal studio apartment in Scandinavian style",
    "resolution": "720x1280",
    "n": 1
  }'

Managed agents

Some providers expose managed agent runtimes instead of only raw model inference.

aihappey-ai treats those as provider capabilities too. The goal is the same: provider-specific agent APIs stay behind normalized gateway contracts where possible.

Examples include provider-native agent surfaces such as Anthropic-managed agents, Z.AI agents and other hosted agent runtimes.

Core MCP servers

aihappey-ai exposes core MCP servers over streamable HTTP.

Discovery:

curl "$BASE_URL/v0.1/servers"

Core MCP server URLs use the same $BASE_URL.

Provider Support Matrix

The table below shows which endpoints each provider implements (✅), not yet implemented (❌), partially implemented (🟡) or for which an endpoint is not applicable to the provider (➖).