Web Search & Thinking

Enable special features like web search, deep thinking, and deep research through request parameters or headers.

Overview

Chat2API supports special features that enhance AI capabilities. These features can be enabled through request parameters or HTTP headers.

Available Features

Feature	Parameter	Header	Description
Web Search	`web_search`	`X-Web-Search`	Enable internet search capability
Deep Thinking	`reasoning_effort`	`X-Reasoning-Effort`	Enable extended reasoning mode
Deep Research	`deep_research`	`X-Deep-Research`	Enable comprehensive research mode

Web Search

Enable AI to search the internet for up-to-date information.

Supported Providers

Provider	Support	Notes
DeepSeek	✅	Fully supported
GLM	✅	Fully supported
Kimi	✅	Fully supported
Qwen	✅	Fully supported

Deep Thinking

Enable extended reasoning for complex problems. The model thinks longer and provides more detailed analysis.

Effort Levels

Level	Description
`low`	Quick thinking, faster response
`medium`	Balanced thinking depth and speed
`high`	Deep thinking, comprehensive analysis

Supported Providers

Provider	Support	Notes
DeepSeek	✅	Fully supported, outputs reasoning process
GLM	✅	Fully supported
Qwen	✅	Fully supported, thinking mode
Qwen AI	✅	Fully supported, outputs reasoning process

Deep Research

Enable comprehensive research mode for in-depth analysis and information gathering.

Supported Providers

Provider	Support	Notes
GLM	✅	Fully supported

Client Configuration

Cline

Configure in VS Code Cline plugin:

Open Cline settings
Set API Base URL to http://localhost:8080/v1
Enter API Key
When setting up model, add in Custom Headers:

Parameter	Type	Value
`web_search`	boolean	`true`
`reasoningEffort`	string	`low` / `medium` / `high`

Roo Code / Kilo

Configure in VS Code Roo Code or Kilo plugin:

Open settings
Set Base URL to http://localhost:8080/v1
Enter API Key
When setting up model, add in Custom Headers:

Parameter	Type	Value
`web_search`	boolean	`true`

Enable reasoning effort and set model reasoning intensity

Cherry Studio

Configure in Cherry Studio:

Open Assistant settings
Add in custom parameters:

Parameter	Type	Value
`web_search`	boolean	`true`
`reasoningEffort`	string	`low` / `medium` / `high`

API Examples

Enable Web Search

curl http://localhost:8080/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer your-api-key" \
  -d '{
    "model": "DeepSeek-V3.2",
    "messages": [
      {"role": "user", "content": "What are the latest AI news today?"}
    ],
    "web_search": true
  }'

Enable Deep Thinking

curl http://localhost:8080/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer your-api-key" \
  -d '{
    "model": "DeepSeek-V3.2",
    "messages": [
      {"role": "user", "content": "Solve this complex math problem step by step..."}
    ],
    "reasoning_effort": "high"
  }'

Combine Features

curl http://localhost:8080/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer your-api-key" \
  -H "X-Web-Search: true" \
  -H "X-Reasoning-Effort: high" \
  -d '{
    "model": "DeepSeek-V3.2",
    "messages": [
      {"role": "user", "content": "Research and analyze the current state of quantum computing..."}
    ]
  }'

Notes

Feature availability depends on provider support, model support, and account status. Verify provider support before using.

Deep thinking and research modes increase response time. Use as needed. Combining features adds further latency.

FAQ

Issue	Solution
Feature not working	Verify provider support, check parameter/header format
Increased response time	Deep thinking and research modes take longer, use lower effort levels
Unexpected results	Features may change model behavior, test with simple queries first

Web Search & Thinking

On this page