AI APIOpenAIAnthropicGoogleLLMDeveloper

The Mainstream AI API Protocols in 2026: Companies, Formats & Working Examples

2026-06-255 Min Read

The Mainstream AI API Protocols in 2026: Companies, Formats & Working Examples

As AI becomes infrastructure, understanding the protocol differences between major vendors is critical for developers. This guide provides technically accurate, production-ready examples of how each API works — no speculation, just real formats.

Part 1: OpenAI (Chat Completions API)

Official Documentation: https://platform.openai.com/docs/api-reference

OpenAI maintains the most widely adopted API format, which has become a de facto standard emulated by many alternative models.

Core Endpoints

Capability Endpoint Method
Text Chat https://api.openai.com/v1/chat/completions POST
Image Generation https://api.openai.com/v1/images/generations POST
Image Understanding (GPT-4o/GPT-4V) https://api.openai.com/v1/chat/completions POST
Audio Transcription (Whisper) https://api.openai.com/v1/audio/transcriptions POST
Text-to-Speech https://api.openai.com/v1/audio/speech POST
Video Generation (Sora, if available) Check current docs —

1. Text Chat Completions (Standard)

Request Example:

POST https://api.openai.com/v1/chat/completions
Authorization: Bearer sk-...your-key...
Content-Type: application/json

{
  "model": "gpt-4o",
  "messages": [
    {
      "role": "system",
      "content": "You are a helpful assistant."
    },
    {
      "role": "user",
      "content": "Explain recursion in one sentence."
    }
  ],
  "temperature": 0.7,
  "max_tokens": 500,
  "stream": false
}

Response Example:

{
  "id": "chatcmpl-...",
  "object": "chat.completion",
  "created": 1719310000,
  "model": "gpt-4o-...",
  "choices": [
    {
      "index": 0,
      "message": {
        "role": "assistant",
        "content": "Recursion is a programming technique where a function calls itself with a modified input until it reaches a base case that stops the recursion."
      },
      "finish_reason": "stop"
    }
  ],
  "usage": {
    "prompt_tokens": 32,
    "completion_tokens": 45,
    "total_tokens": 77
  },
  "system_fingerprint": "fp_..."
}

2. Multimodal (Image Input)

Request Example:

POST https://api.openai.com/v1/chat/completions
Authorization: Bearer sk-...your-key...
Content-Type: application/json

{
  "model": "gpt-4o",
  "messages": [
    {
      "role": "user",
      "content": [
        {
          "type": "text",
          "text": "What's in this image?"
        },
        {
          "type": "image_url",
          "image_url": {
            "url": "https://example.com/image.jpg",
            "detail": "high"
          }
        }
      ]
    }
  ],
  "max_tokens": 300
}

Or base64 inline:

POST https://api.openai.com/v1/chat/completions
Authorization: Bearer sk-...your-key...
Content-Type: application/json

{
  "model": "gpt-4o",
  "messages": [
    {
      "role": "user",
      "content": [
        {
          "type": "text",
          "text": "Describe this image."
        },
        {
          "type": "image_url",
          "image_url": {
            "url": "data:image/jpeg;base64,/9j/4AAQSkZJRg..."
          }
        }
      ]
    }
  ]
}

Response Structure: Same as text chat (see above).

3. Audio Transcription (Whisper)

Request Example (multipart/form-data):

POST https://api.openai.com/v1/audio/transcriptions
Authorization: Bearer sk-...your-key...
Content-Type: multipart/form-data; boundary=----boundary

------boundary
Content-Disposition: form-data; name="file"; filename="audio.mp3"
Content-Type: audio/mpeg

[...binary audio data...]
------boundary
Content-Disposition: form-data; name="model"

whisper-1
------boundary
Content-Disposition: form-data; name="response_format"

json
------boundary--

Response Example:

{
  "text": "The quick brown fox jumps over the lazy dog."
}

Or verbose JSON:

{
  "task": "transcribe",
  "language": "english",
  "duration": 2.5,
  "text": "The quick brown fox jumps over the lazy dog.",
  "segments": [
    {
      "id": 0,
      "seek": 0,
      "start": 0.0,
      "end": 2.5,
      "text": "The quick brown fox jumps over the lazy dog."
    }
  ]
}

4. Text-to-Speech

Request Example:

POST https://api.openai.com/v1/audio/speech
Authorization: Bearer sk-...your-key...
Content-Type: application/json

{
  "model": "tts-1",
  "input": "The quick brown fox jumps over the lazy dog.",
  "voice": "alloy",
  "response_format": "mp3",
  "speed": 1.0
}

Response: Binary audio stream (MP3).

5. Image Generation (DALL-E)

Request Example:

POST https://api.openai.com/v1/images/generations
Authorization: Bearer sk-...your-key...
Content-Type: application/json

{
  "model": "dall-e-3",
  "prompt": "A cute robot cat sitting on a windowsill, sunset in background, cyberpunk style.",
  "n": 1,
  "size": "1024x1024",
  "quality": "standard",
  "response_format": "url"
}

Response Example:

{
  "created": 1719310000,
  "data": [
    {
      "url": "https://oaidalleapiprodscus.blob.core.windows.net/...",
      "revised_prompt": "A cute robotic cat with metallic fur sitting elegantly on a windowsill..."
    }
  ]
}

Part 2: Anthropic (Claude Messages API)

Official Documentation: https://docs.anthropic.com/claude/reference/messages_post

Anthropic uses a clean, thoughtfully designed API format focused on safety and long context.

Core Endpoints

Capability Endpoint Method
Text & Multimodal Chat https://api.anthropic.com/v1/messages POST

1. Text Messages (Standard)

Request Example:

POST https://api.anthropic.com/v1/messages
x-api-key: sk-ant-...your-key...
anthropic-version: 2023-06-01
Content-Type: application/json

{
  "model": "claude-3-5-sonnet-20241022",
  "max_tokens": 1024,
  "messages": [
    {
      "role": "user",
      "content": "Explain recursion in one sentence."
    }
  ],
  "system": "You are a helpful, precise assistant.",
  "temperature": 0.7
}

Response Example:

{
  "id": "msg_...",
  "type": "message",
  "role": "assistant",
  "model": "claude-3-5-sonnet-20241022",
  "content": [
    {
      "type": "text",
      "text": "Recursion is a programming technique where a function calls itself with progressively simpler inputs until it reaches a base case that can be solved directly."
    }
  ],
  "stop_reason": "end_turn",
  "stop_sequence": null,
  "usage": {
    "input_tokens": 25,
    "output_tokens": 48
  }
}

2. Multimodal (Images)

Request Example:

POST https://api.anthropic.com/v1/messages
x-api-key: sk-ant-...your-key...
anthropic-version: 2023-06-01
Content-Type: application/json

{
  "model": "claude-3-5-sonnet-20241022",
  "max_tokens": 1024,
  "messages": [
    {
      "role": "user",
      "content": [
        {
          "type": "image",
          "source": {
            "type": "base64",
            "media_type": "image/jpeg",
            "data": "/9j/4AAQSkZJRg..."
          }
        },
        {
          "type": "text",
          "text": "What's in this image?"
        }
      ]
    }
  ]
}

Response Structure: Same as text messages (see above).

3. Multi-Turn Conversation

Request Example:

POST https://api.anthropic.com/v1/messages
x-api-key: sk-ant-...your-key...
anthropic-version: 2023-06-01
Content-Type: application/json

{
  "model": "claude-3-5-sonnet-20241022",
  "max_tokens": 1024,
  "messages": [
    {
      "role": "user",
      "content": "What is the capital of France?"
    },
    {
      "role": "assistant",
      "content": "The capital of France is Paris."
    },
    {
      "role": "user",
      "content": "And how far is it from London?"
    }
  ]
}

Response Example:

{
  "id": "msg_...",
  "type": "message",
  "role": "assistant",
  "model": "claude-3-5-sonnet-20241022",
  "content": [
    {
      "type": "text",
      "text": "Paris is approximately 344 kilometers (214 miles) from London by direct air travel."
    }
  ],
  "stop_reason": "end_turn",
  "usage": {
    "input_tokens": 68,
    "output_tokens": 32
  }
}

Part 3: Google (Gemini API)

Official Documentation: https://ai.google.dev/api

Google's Gemini API supports unified multimodal interactions across text, image, audio, and video.

Core Endpoints

Capability Endpoint Method
Text & Multimodal (Generate Content) https://generativelanguage.googleapis.com/v1beta/models/{model}:generateContent POST
Streamed Generation https://generativelanguage.googleapis.com/v1beta/models/{model}:streamGenerateContent POST

Note: Replace {model} with gemini-2.0-flash-exp, gemini-1.5-pro, etc.

1. Text Only (Generate Content)

Request Example:

POST https://generativelanguage.googleapis.com/v1beta/models/gemini-1.5-pro:generateContent?key=YOUR_API_KEY
Content-Type: application/json

{
  "contents": [
    {
      "role": "user",
      "parts": [
        {
          "text": "Explain recursion in one sentence."
        }
      ]
    }
  ],
  "generationConfig": {
    "temperature": 0.7,
    "maxOutputTokens": 500
  }
}

Response Example:

{
  "candidates": [
    {
      "content": {
        "parts": [
          {
            "text": "Recursion is a programming technique where a function calls itself with a modified, smaller problem until it reaches a base case that can be solved directly, unwinding the call stack to produce the final result."
          }
        ],
        "role": "model"
      },
      "finishReason": "STOP"
    }
  ],
  "usageMetadata": {
    "promptTokenCount": 12,
    "candidatesTokenCount": 52,
    "totalTokenCount": 64
  }
}

2. Multimodal (Text + Image)

Request Example:

POST https://generativelanguage.googleapis.com/v1beta/models/gemini-1.5-pro:generateContent?key=YOUR_API_KEY
Content-Type: application/json

{
  "contents": [
    {
      "role": "user",
      "parts": [
        {
          "inline_data": {
            "mime_type": "image/jpeg",
            "data": "/9j/4AAQSkZJRg..."
          }
        },
        {
          "text": "What's in this image?"
        }
      ]
    }
  ]
}

Response Structure: Same as text generation (see above).

3. Multimodal (Text + Audio)

Request Example:

POST https://generativelanguage.googleapis.com/v1beta/models/gemini-1.5-pro:generateContent?key=YOUR_API_KEY
Content-Type: application/json

{
  "contents": [
    {
      "role": "user",
      "parts": [
        {
          "inline_data": {
            "mime_type": "audio/mp3",
            "data": "SUQzBAAAAAAAI1RTU0UAAAAPAAADTGF2ZjU2..."
          }
        },
        {
          "text": "Transcribe this audio."
        }
      ]
    }
  ]
}

Response Example:

{
  "candidates": [
    {
      "content": {
        "parts": [
          {
            "text": "The quick brown fox jumps over the lazy dog."
          }
        ],
        "role": "model"
      }
    }
  ],
  "usageMetadata": {
    "promptTokenCount": 842,
    "candidatesTokenCount": 18,
    "totalTokenCount": 860
  }
}

4. Multimodal (Text + Video)

Request Example:

POST https://generativelanguage.googleapis.com/v1beta/models/gemini-1.5-pro:generateContent?key=YOUR_API_KEY
Content-Type: application/json

{
  "contents": [
    {
      "role": "user",
      "parts": [
        {
          "inline_data": {
            "mime_type": "video/mp4",
            "data": "AAAAIGZ0eXBtcDQyAAAAAG1wNDJpc29tAAACAG1wNDJpc3..."
          }
        },
        {
          "text": "Summarize what happens in this video."
        }
      ]
    }
  ]
}

Response Example:

{
  "candidates": [
    {
      "content": {
        "parts": [
          {
            "text": "The video shows a cat sitting on a windowsill, watching birds outside. After about 10 seconds, the cat stands up and moves away from the window."
          }
        ],
        "role": "model"
      }
    }
  ],
  "usageMetadata": {
    "promptTokenCount": 2846,
    "candidatesTokenCount": 48,
    "totalTokenCount": 2894
  }
}

Part 4: Quick Comparison Cheat Sheet

Aspect OpenAI Anthropic Google Gemini
Authentication Authorization: Bearer sk-... x-api-key: sk-ant-... + anthropic-version URL query ?key=...
Text Message Key messages[].content messages[].content contents[].parts[].text
System Prompt { role: "system", ... } Top-level "system" field systemInstruction field
Image Input { type: "image_url", ... } { type: "image", source: {...} } { inline_data: { mime_type, data } }
Streaming Flag stream: true (check docs) Use :streamGenerateContent
Usage Field usage: { prompt_tokens, completion_tokens } usage: { input_tokens, output_tokens } usageMetadata: { promptTokenCount, ... }

Part 5: cURL Quick Reference (Production-Ready)

OpenAI (cURL)

curl https://api.openai.com/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $OPENAI_API_KEY" \
  -d '{
    "model": "gpt-4o",
    "messages": [{"role": "user", "content": "Hello!"}]
  }'

Anthropic (cURL)

curl https://api.anthropic.com/v1/messages \
  -H "Content-Type: application/json" \
  -H "x-api-key: $ANTHROPIC_API_KEY" \
  -H "anthropic-version: 2023-06-01" \
  -d '{
    "model": "claude-3-5-sonnet-20241022",
    "max_tokens": 1024,
    "messages": [{"role": "user", "content": "Hello!"}]
  }'

Google Gemini (cURL)

curl "https://generativelanguage.googleapis.com/v1beta/models/gemini-1.5-pro:generateContent?key=$GEMINI_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "contents": [{"role": "user", "parts": [{"text": "Hello!"}]}]
  }'

Important Notes

This guide reflects current stable API patterns as of mid-2026. Always refer to the official documentation before deploying to production, as APIs evolve.

For checking how AI-visible your own site is, use our AI Visibility Checker tool.

Related Tools

AI Visibility Checker

Check your website's visibility in ChatGPT, Claude, and Perplexity search indexes.

Schema Audit Tool

Identify missing JSON-LD schemas and FAQ structures essential for AI Search visibility.