Text and chat

Build text generation and multi-turn conversations with Chat Completions or Responses.

Overview

AGTCloud provides text generation and multi-turn chat capabilities. After selecting a model, send requests through Chat Completions, Responses, or another text protocol supported by that model.

Core capabilities

Text generation: Generate, rewrite, or expand content from instructions.
Multi-turn conversations: Continue context through message history or response state.
Summarization and classification: Extract key points or organize content into specified categories.
Information extraction: Extract required fields and conclusions from input text.
Advanced capabilities: Support for streaming, tool calling, and structured output depends on the selected model and protocol.

Featured models

Model	Provider	Highlights
DeepSeek	DeepSeek	A strong fit for general chat, reasoning, and coding tasks
Qwen	Alibaba Cloud	Designed for multilingual generation and broad content workflows
Kimi	Moonshot AI	Well suited to long-context understanding, summaries, and conversations
GLM	Zhipu AI	Useful for bilingual chat, writing, and information extraction

View all models

API access

Base URL

https://api.agtcloud.cn/v1

Chat Completions

curl 'https://api.agtcloud.cn/v1/chat/completions' \
  -H 'Content-Type: application/json' \
  -H 'Authorization: Bearer your-api-key' \
  --data '{
    "model": "MODEL_ID",
    "messages": [
      {"role": "system", "content": "Keep the answer concise."},
      {"role": "user", "content": "Describe AGTCloud in one sentence."}
    ]
  }'

Responses

curl 'https://api.agtcloud.cn/v1/responses' \
  -H 'Content-Type: application/json' \
  -H 'Authorization: Bearer your-api-key' \
  --data '{
    "model": "MODEL_ID",
    "input": "Describe AGTCloud in one sentence."
  }'

Common parameters

Parameter	Type	Description
`model`	string	Model ID used by either endpoint
`messages`	array	Message history for Chat Completions
`input`	string / array	Text or input items for Responses
`stream`	boolean	Enables streaming output
`temperature`	number	Adjusts generation randomness when supported by the model
`previous_response_id`	string	Continues an earlier Responses interaction

Chat Completions uses messages; Responses uses input. See the corresponding API Reference for complete parameters. To compare Anthropic Messages, Gemini Generate Content, and other options, see Choose a protocol.