Claude+qwen3-vl:8B HTTP交互信息解释

Claude Code的控制台客户端安装

复制代码

npm install -g @anthropic-ai/claude-code

Ollama的安装，并拉取qwen3-vl:8B的模型

复制代码

ollama pull qwen3-vl:8B

ollama作为service启动后，默认只开启在127.0.0.1端口上，需要修改服务配置达成到0.0.0.0

打开控制台，修改环境变量，而后启动claude

复制代码

export ANTHROPIC_AUTH_TOKEN=ollama
export ANTHROPIC_BASE_URL=http://192.168.1.190:11434 //本地安装的ollama服务
claude

启动后，会和http服务器进行一次http交互，只是确认服务的存在。

进入控制台后选择模型

复制代码

/model qwen3-vl:8B

----------HTTPRequest-----------
POST / v1 / messages ? beta = true HTTP / 1.1
host: localhost: 11434
connection: keep - alive
Accept: application / json
X - Stainless - Retry - Count: 0
X - Stainless - Timeout: 600
X - Stainless - Lang: js
X - Stainless - Package - Version: 0.81 .0
X - Stainless - OS: MacOS
X - Stainless - Arch: x64
X - Stainless - Runtime: node
X - Stainless - Runtime - Version: v22 .4 .1
anthropic - dangerous - direct - browser - access: true
anthropic - version: 2023 - 06 - 01
x - app: cli
User - Agent: claude - cli / 2.1 .110(external, cli)
X - Claude - Code - Session - Id: 745 a401e - a7a4 - 4 aba - 830e- b63bf409bedb
Authorization: Bearer ollama
content - type: application / json
anthropic - beta: claude - code - 20250219, interleaved - thinking - 2025 - 05 - 14, redact - thinking - 2026 - 02 - 12, context - management - 2025 - 06 - 27, prompt - caching - scope - 2026 - 01 - 05
accept - language: *
	sec - fetch - mode: cors
accept - encoding: gzip, deflate
content - length: 533

{
	"model": "qwen3-vl:8B",
	"max_tokens": 1,
	"system": [{
		"type": "text",
		"text": "x-anthropic-billing-header: cc_version=2.1.110.610; cc_entrypoint=cli; cch=00000;"
	}, {
		"type": "text",
		"text": "You are Claude Code, Anthropic's official CLI for Claude."
	}],
	"messages": [{
		"role": "user",
		"content": [{
			"type": "text",
			"text": "Hi",
			"cache_control": {
				"type": "ephemeral"
			}
		}]
	}],
	"metadata": {
		"user_id": "{\"device_id\":\"b1581fe7235e6eed0f44c26a8c26d7fc1c704ce894c977cb770f421cf7720923\",\"account_uuid\":\"\",\"session_id\":\"745a401e-a7a4-4aba-830e-b63bf409bedb\"}"
	}
}


----------HTTPResponse-----------
HTTP / 1.1 200 OK
Content - Type: application / json
Date: Fri, 17 Apr 2026 07: 31: 04 GMT
Content - Length: 219

{
	"id": "msg_c1a7f4c4be5ae8ca4843bd20",
	"type": "message",
	"role": "assistant",
	"model": "qwen3-vl:8B",
	"content": [{
		"type": "thinking",
		"thinking": "Okay"
	}],
	"stop_reason": "max_tokens",
	"usage": {
		"input_tokens": 67,
		"output_tokens": 1
	}
}

确认大模型端对qwen3-vl:8B的支持，此时也会该Session创建了唯一性的session编号。

提问: 创建文件index.html

两个HTTP交互见附件

claude向qwen3-vl发起问题，提交<system-reminder>的提示词，模型指定命令<command-name>/model</command-name>，用户问题

{

"type": "text",

"text": "创建文件index.html",

"cache_control": {

"type": "ephemeral"

}

}，通过tools":表达技能能力。qwen进行llm分析后，最终会回复tool use的调用

data: {

"type": "content_block_start",

"index": 0,

"content_block": {

"type": "tool_use",

"id": "call_ooye05av",

"name": "Write",

"input": {}

}

本地调用工具执行后，claude再次发起交互，汇报工具调用结果，

"content": [{

"tool_use_id": "call_ooye05av",

"type": "tool_result",

"content": "File created successfully at: /Users/xuxin/index.html",

"cache_control": {

"type": "ephemeral"

}

}] qwen返回结果描述以及后续建议。

如有部署疑问和私信沟通。也可浏览agent-llm概念文章。