Text Generation Overview

APIFree provides a unified text generation layer across multiple protocol styles while keeping authentication, billing, and model access consistent at the platform level.

Available endpoints

Endpoint	Protocol style	Streaming	Best for
POST /v1/chat/completions	OpenAI Chat Completions	Yes	Most chat-based LLM integrations and existing OpenAI-compatible clients
POST /v1/messages	Anthropic Messages	Yes	Applications already built around Anthropic-style message payloads
POST /v1/responses	OpenAI Responses	Yes	New OpenAI-style text generation integrations
POST /google/gemini	Google Gemini	Yes	Clients that need Gemini-native request and response structures

How to choose

• Use POST /v1/chat/completions if your application already speaks the classic OpenAI chat format with messages, or if you want the most familiar migration path.
• Use POST /v1/messages if your application already sends messages in Anthropic’s format and expects Anthropic-style events.
• Use POST /v1/responses if you want an OpenAI Responses-compatible interface for text generation and tool orchestration.
• Use POST /google/gemini if you need Gemini-native protocol compatibility instead of OpenAI- or Anthropic-style payloads.

Common behavior

All text generation endpoints:

• Use the same APIFree API key authentication model.
• Require a model field in the request body.
• Validate the request against the selected model schema before forwarding it upstream.
• Support standard JSON responses and streaming responses when stream is enabled.

Authentication

Include your API key in the Authorization header:

Authorization: Bearer YOUR_API_KEY

Streaming

When stream is true, the API returns the upstream stream format for the selected protocol:

• chat/completions: OpenAI Chat Completions SSE chunks
• messages: Anthropic-style SSE events
• responses: OpenAI-style SSE events
• google/gemini: Gemini stream output format

Choose the endpoint whose request and stream shape best matches your client runtime.