App profile
OpenClaw
By OpenClaw
A self-hosted personal AI assistant with custom model providers. The site text API is open; a separate client test is pending.
GPT · Chat Completions
Partial acceptance
The real read tool and exact random-file reply passed. The initial exact reply omitted its opening characters, so overall acceptance is partial and the cause remains unresolved.
API status · verifiedThe protocol used by this client succeeded in its corresponding GPT client test.
- Provider mode
- OpenAI-compatible
- Key service
- GPT
- Client base URL
https://api.openoxai.com/v1- Request endpoint
POST https://api.openoxai.com/v1/chat/completions- Model ID
gpt-6.1-sol
- 1
Prepare a GPT service key
Select the GPT service in Keys and confirm gpt-6.1-sol is allowed for that key.
- 2
Configure the custom provider
Select OpenAI-compatible and enter the URL and model below. Set the API key environment variable in the terminal that launches the client. Merge the configuration while preserving other providers.
- 3
Check the first request
Send a simple message and check its model, completion and cost in Usage. Tool support and charges depend on the selected model.
export OPENOX_API_KEY="YOUR_OPENOX_API_KEY"This applies to the current terminal session. Replace the placeholder and launch the client from the same terminal. Do not put the real key in the config file.
{
"models": {
"providers": {
"openox": {
"baseUrl": "https://api.openoxai.com/v1",
"apiKey": "${OPENOX_API_KEY}",
"api": "openai-completions",
"agentRuntime": {
"id": "openclaw"
},
"models": [
{
"id": "gpt-6.1-sol",
"name": "gpt-6.1-sol",
"input": [
"text"
]
}
]
}
}
},
"agents": {
"defaults": {
"model": {
"primary": "openox/gpt-6.1-sol"
}
}
}
}Replace the placeholder with your own key. CLI environment variables use macOS, Linux or WSL shell syntax. Preserve existing settings.