Skip to main content
Qwen Cloud home page
Docs
Clients & tools

Chatbox

Zero-setup AI chat client

Chatbox is a cross-platform desktop AI chat application that supports text, images, and documents. Connect it directly to Qwen Cloud's pay-as-you-go API for cost-effective access to all Qwen models through the OpenAI-compatible endpoint.

Quick start

Get running in a few minutes: 
# 1. Install
Download from chatboxai.app or use web version

# 2. Configure (open Settings > Model Provider > click Add)
Name: QwenCloud
API Mode: Select "OpenAI API Compatible"
API Key: sk-xxx
API Host: https://dashscope-intl.aliyuncs.com/compatible-mode/v1

# 3. Test (add a model and chat)
Click "New" in Model section
Enter Model ID: qwen3.5-plus
Type a message: "What is machine learning?"
You should see: qwen3.5-plus provides a detailed explanation of machine learning

Configuration

Basic setup

Configure Chatbox to use Qwen Cloud:
  • API endpoint: https://dashscope-intl.aliyuncs.com/compatible-mode/v1
  • Authentication: API key required
  • Model selection: Add any Qwen model by entering its ID

Step-by-step configuration

1

Add custom provider

Settings → Model Provider → Add
  • Name: QwenCloud
  • API Mode: OpenAI API Compatible
2

Configure API

  • API Key: Your API key
  • API Host: https://dashscope-intl.aliyuncs.com/compatible-mode/v1
  • Leave API Path empty
3

Add models

In Model section, click New:
  • Model ID: Enter model name (e.g., qwen3.5-plus)
  • Select capabilities: Reasoning, Tool use, Vision (as applicable)
4

Optimize chat settings

Chat Settings:
  • Max Message Count in Context: 5-10 (for casual chat)
  • Temperature: 0.1-0.9 (lower = more focused)
  • Top P: ≤1.0

Model capabilities

When adding models, select appropriate capabilities:
CapabilityUse forExample models
ReasoningThinking modeqwen3-max, qwq-32b-preview
Tool useFunction callingMost Qwen models
VisionImage understandingqwen3-vl-plus, qvq-72b-preview

Limitations

  • Does not support: Audio/video files in chat
  • Document parsing: Cannot extract images from documents
  • Context limits: Multi-turn conversations accumulate tokens quickly

Examples

  • Text conversation
  • Image understanding
  • Document analysis
You: Explain quantum computing in simple terms
Model: qwen3.5-plus
[Chatbox shows thinking process if reasoning model is used]

Troubleshooting

"Failed to connect to Custom Provider"
Solution: Verify API key and endpoint are correct. Check quota in Qwen Cloud.
"Range of input length should be [1, xxx]"
Solution:
  • First turn: Input too long, use model with larger context
  • Multi-turn: Reduce "Max Message Count in Context" or start new chat
  • Switch to 1M context model for long conversations
"'temperature' must be Float"
Solution: Set Temperature to less than 2.0 in Chat Settings
"xx is greater than the maximum of 1 - 'top_p'"
Solution: Set Top P to 1.0 or less in Chat Settings
High token consumption
Solution:
  • Lower "Max Message Count in Context" to 5-10
  • Start new chats for unrelated topics
  • Use qwen3.5-flash with context cache for document Q&A

Cost optimization

Chatbox conversations can consume many tokens due to:
  • Multi-turn context accumulation
  • Document parsing
  • Image processing
Reduce costs by:
  1. Context management: Set Max Message Count to 5-10
  2. Model selection: Use qwen3.5-flash for routine tasks
  3. Context cache: Models like qwen-flash support caching for repeated content
  4. New chats: Start fresh for unrelated topics