LobeChat integration

Use LobeChat with Chinese AI models through ChinaAPI.

Set LobeChat's OpenAI provider to ChinaAPI's proxy URL and chat with Chinese model families using one OpenAI-compatible key.

Start free — $2 credit Live model pricing

Prerequisites

Before configuring LobeChat.

LobeChat access

Use LobeChat Cloud or your self-hosted LobeChat instance with provider settings enabled.

ChinaAPI key

Create an sk-... token in Dashboard -> Console -> Tokens.

Setup

Configure the OpenAI provider.

Open LobeChat settings, then go to AI Providers.
Locate the OpenAI provider section and enable it if needed.
Paste your ChinaAPI token into the OpenAI API key field.
Set the OpenAI proxy URL, endpoint, or custom API base field to https://api.chinaapi.ai/v1. The exact label varies between Cloud, desktop, and self-hosted builds.
Add model IDs manually if your LobeChat build does not fetch them from the provider. Start with deepseek-flash, kimi-k2.6, or qwen3.7-plus.
Create a new chat and select one of the ChinaAPI model IDs.

Copy-paste test

Use a short in-tool prompt.

Reply in one sentence: LobeChat is using ChinaAPI for Chinese models.

If the in-tool request fails, verify the key with curl:

curl https://api.chinaapi.ai/v1/chat/completions \
  -H "Authorization: Bearer sk-..." \
  -H "Content-Type: application/json" \
  -d '{"model":"qwen3.7-plus","messages":[{"role":"user","content":"Say LobeChat is connected."}],"stream":true}'

Recommended models

Good LobeChat defaults.

For current model details and rate evidence, review DeepSeek-V4 Flash and Qwen3.7 Plus before choosing a production default.

`deepseek-flash`

Fast chat, coding help, and general assistant usage.

`qwen3.7-plus`

Balanced reasoning, planning, and multimodal-friendly work.

`kimi-k2.6`

General Kimi model for long chat context and planning.

`glm-5.1`

Strong GLM option for reasoning, coding, and agent-style chat.

`MiniMax-M2.7`

Efficient MiniMax model for lower-latency chat workflows.

Troubleshooting

Common LobeChat issues.

Thinking models such as qwen3.7-max can generate reasoning tokens before visible output. Enable streaming if your LobeChat build exposes it, or raise proxy/client timeout settings in self-hosted deployments.

Model not found: check live pricing and enter the exact ID in LobeChat's model list.

401: make sure the API key starts with sk-, is saved under the OpenAI provider, and is paired with the ChinaAPI proxy URL.

Pricing and signup

Try Chinese chat models with $2 free credit.

ChinaAPI bills token usage in USD. Standard model rates follow providers' international list prices at 100%; some models have a separate Paid rate after the first top-up. Check the live pricing page for exact current IDs and availability.