OpenHands
OpenHands (formerly OpenDevin) is an open-source, AI-driven software development platform. It supports one-click deployment via Docker, offers both Web UI and CLI interaction modes, and can autonomously complete development tasks such as code writing, debugging, and terminal operations.
Prerequisites
- You have activated a TCB environment and obtained the environment ID
cloudBaseEnvID - Enable the required model in the AI console
- Obtain the
Base URLandAPI Keyin the AI console

Configuration Steps
OpenHands supports connecting custom models through Docker environment variables or the config.toml configuration file.
Configuration Field Mapping Table
| Field | Explanation |
|---|---|
| model | The large model ID. Copy the identifier of an enabled large model from the model list in the AI console and enter it here. |
| base_url | The Base URL for the large model. Obtain it from the AI console. |
| api_key | The API Key for calling the large model, which is obtained in the AI console. |
- Configuration During Docker Startup
- config.toml Configuration File
Pass the environment variables through the -e parameter:
docker run -it --rm --pull=always \
-e LLM_MODEL="hy3-preview" \
-e LLM_API_KEY="eyJhbGciOiJSUzI1N..." \
-e LLM_BASE_URL="https://{{cloudBaseEnvID}}.api.tcloudbasegateway.com/v1/ai/cloudbase" \
-p 3000:3000 \
--add-host host.docker.internal:host-gateway \
-v /var/run/docker.sock:/var/run/docker.sock \
ghcr.io/openhandsai/openhands:main
After startup, access http://localhost:3000 to use it.
note
This method is available only in development mode (started via main.py or cli.py). For the Docker method, use the environment variables from the "Configuration During Docker Startup" tab.
- Open or create the
config.tomlconfiguration file.
~/.openhands/config.toml
- Enter the large model configuration in the
[llm]field.
[llm]
model = "hy3-preview"
api_key = "eyJhbGciOiJSUzI1N..."
base_url = "https://{{cloudBaseEnvID}}.api.tcloudbasegateway.com/v1/ai/cloudbase"
- Start up OpenHands to use the configured custom model.
Common Error Troubleshooting
| Error Message | Solution |
|---|---|
404 Model Not Found | The Base URL is entered incorrectly. Go to the AI console to copy the correct Base URL. |
401 Authentication Failed | The API Key is entered incorrectly or may have been deleted. |
429 Token usage exceeded quota | An incorrect large model name is configured, or the large model is not enabled in the AI console. |