Can I use LanoX with Codex or Claude Code?

Yes. For Codex, use https://api.lanox.ai/v1 and configure the Responses API as described in the documentation. For Claude Code, use https://api.lanox.ai without /v1. Both require a LanoX API key and a model ID that supports the client's API format; follow the LanoX CC Switch guide for setup instructions.

How do I check my LanoX model settings and request logs?

Copy the exact model ID from the current catalog, then compare it with the model parameter sent by your client, the model information in the API response, and the request details in your dashboard. Similar names or different versions should not be treated as the same model. If the records do not match, contact support with the request time and request ID so the call can be investigated.

What do I need to change when switching from a provider's API to LanoX?

If LanoX supports the API format your project uses, you can usually keep your SDK and update the base URL, API key, and model ID. Test the features your application relies on, including streaming, tool calls, and image inputs. Check compatibility separately for Chat Completions, Responses, and Claude Messages.

What's the difference between a compatible API and a model's native API?

A model's native API is defined by its provider and may include provider-specific features and parameters. A compatible API matches the request and response format expected by existing clients, making those clients easier to reuse. Support for streaming, tool calls, structured output, and multimodal inputs can differ, so check each feature your application needs.

OPENAI-COMPATIBLE API & CLIENT SETUP

LanoX docs for developersfor the AI models you need

The LanoX docs hub helps you connect OpenAI, Claude, Grok, DeepSeek, and other models to your app, Codex, or Claude Code. Add your API key and choose an exact model ID from the live catalog — everything you need is in one documentation hub.

SOFTWARE

Connect LanoX to the software you use

Codex

OpenAI coding assistant

Setup guide

Claude Code

Anthropic coding assistant

Setup guide

WorkBuddy

AI work assistant

Setup guide

DeepSeek Harness

Flexible model calling tool

Setup guide

Already have an API key? Start configuring

SKILLS

Add creative skills to your software

LanoX Image

lanox-image

Codex

Generate and edit images with your choice of model.

Text to imageReference image editing

Default model: gpt-image

LanoX Video

lanox-video

Codex

Generate videos with a specified model and track task progress.

Video generationTask tracking

Default model: seedance

Skills are optional and independent of software setup. Model calls are billed by actual usage.

View pricing

FOR DEVELOPERS

Connect directly through the API

API address Hosthttps://api.lanox.ai

Current model3 available endpoints

OpenAIBearer Token

Chat Completions

Returns conversational model output synchronously or as an SSE stream.

POST/v1/chat/completions

Model ID: gpt-5.6-sol
Authentication: Authorization: Bearer $LANOX_API_KEY

Request parameters

application/json

Field	Type	Required	Description
`model`	string	Required	The model identifier, such as gpt-5.6-sol. The gateway requires a non-empty value.
`messages`	object[]	Optional	An OpenAI-compatible message array. The gateway forwards it unchanged; the selected model determines whether it is required and which message types are supported.
`messages[].role`	developer \| system \| user \| assistant \| tool	Optional	The message role.
`messages[].content`	string \| object[] \| null	Optional	Plain text or an array of multimodal content blocks. It may be null for an assistant message that contains only tool calls.
`messages[].content[].type`	text \| input_text \| image_url \| string	Optional	The content-block type. Available types depend on the selected model.
`messages[].content[].text`	string	Optional	The text carried by a text or input_text content block.
`messages[].content[].image_url`	string \| object	Optional	An image URL accepted by the compatible model, either as a string or as an object containing url and detail.
`messages[].name`	string	Optional	An optional participant name.

Response parameters

Field	Type	Required	Description
`id`	string	Optional	A common Chat Completion response ID. The gateway neither rewrites nor guarantees this field.
`object`	"chat.completion"	Optional	The object type for a non-streaming response.
`created`	integer	Optional	The creation time as a Unix timestamp in seconds.
`model`	string	Optional	The model that generated the response.
`choices`	object[]	Optional	The array of completion candidates.
`choices[].index`	integer	Optional	The candidate index.
`choices[].message`	object	Optional	The assistant message returned by a non-streaming request.
`choices[].message.role`	"assistant"	Optional	The assistant message role.

The Java gateway requires only a non-empty valid JSON body and a non-empty model. It forwards messages, tools, response_format, and sampling parameters unchanged; exact support depends on the selected model.
Successful responses pass through the downstream status, headers, and body without a business response envelope. The table documents a common OpenAI-compatible shape and does not imply that the gateway adds missing fields.
The image_generation tool is not supported on this endpoint. Use /v1/responses instead.
A disconnected stream does not mean that the request was not executed. Evaluate the risk of a duplicate request before retrying.

Request example

curl -X POST "https://api.lanox.ai/v1/chat/completions" \
  -H "Authorization: Bearer $LANOX_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "model": "gpt-5.6-sol",
  "messages": [
    {
      "role": "user",
      "content": [
        {
          "type": "text",
          "text": "Check the weather in Shanghai"
        }
      ]
    }
  ],
  "tools": [
    {
      "type": "function",
      "function": {
        "name": "get_weather",
        "description": "Look up the weather for a city",
        "parameters": {
          "type": "object",
          "properties": {
            "city": {
              "type": "string"
            }
          },
          "required": [
            "city"
          ],
          "additionalProperties": false
        },
        "strict": true
      }
    }
  ],
  "tool_choice": "auto",
  "temperature": 0.2,
  "max_completion_tokens": 300,
  "stream": false
}'

Response example

{
  "id": "chatcmpl_example",
  "object": "chat.completion",
  "created": 1785196800,
  "model": "gpt-5.6-sol",
  "choices": [{
    "index": 0,
    "message": {
      "role": "assistant",
      "content": null,
      "tool_calls": [{
        "id": "call_weather_01",
        "type": "function",
        "function": {"name": "get_weather", "arguments": "{\"city\":\"Shanghai\"}"}
      }]
    },
    "finish_reason": "tool_calls"
  }],
  "usage": {
    "prompt_tokens": 42,
    "completion_tokens": 18,
    "total_tokens": 60,
    "prompt_tokens_details": {"cached_tokens": 0}
  }
}

Read your API key from a local environment variable

Reference documentation only. Check GET /v1/models for model access and available endpoints.View models and pricing

After your first call, check these 4 items

Requested model
Actual result
Usage record
Final charge

Usage details

DOCUMENTATION

LanoX Documentation

API reference, software setup and your first call, all in one place.

API reference

Request examples, parameters and response fields for text, image and video APIs.

Read API documentation

Client setup guides

Addresses and configuration checks for Codex, Claude Code and CC Switch.

View client guides

Quick start and verification

Create an API key and verify models, results, usage and charges.

Create an API key

Image API · lanox-image-2

Text to image, image editing, task creation and result retrieval.

View dedicated guide

FAQ

Frequently asked questions

Answers about your account, credits, and API setup.

Accounts and service

Could this put my account at risk?

LanoX connects to model services through an API key, separately from your personal ChatGPT or Claude subscription. You still need to follow the platform's and model provider's usage policies. This setup is not a guarantee of zero risk.

Which models can I use?

Use models from GPT, Claude, Gemini, DeepSeek, Qwen, and other families with a single LanoX balance. For current availability, features, and rates, check Models & pricing. To connect your app, follow the setup guides.

How does LanoX keep prices low?

Enterprise purchasing and dynamic resource routing help lower costs. You pay for usage at LanoX's published model rates, rather than buying a separate subscription for each model. Discounts and rates vary by model; see the published pricing for details.

Do my credits expire?

Standard top-up credits have no expiry date while your account remains open. Charges are deducted as you use the service.

Time-limited credits, such as daily or weekly passes, are an exception. Check the terms for the product you purchase.

How is LanoX different from a Claude subscription?

LanoX provides API access to models, not a subscription to the Claude website. Add your LanoX API key to a compatible app and pay for usage from one shared balance. Features and limits depend on the model and app you choose.

Setup and API requests

Do Codex and Claude Code use the same base URL?

No. For Codex, use https://api.lanox.ai/v1 with the native Responses format. For Claude Code, use https://api.lanox.ai without /v1. Follow the setup guide for your app.

Do I need to install a skill after connecting an app?

No. App setup and skills are separate. Add a skill only when you want image or video creation: install it in a Codex chat, add your API key, then restart Codex.

Which model name should I use?

Copy the exact model ID from the catalog or GET /v1/models . Make sure it supports your endpoint. Use the ID, not the model's display name.

Why didn't my image or video task return a result right away?

A successful creation response means the task was accepted, not finished. Save the task ID and poll its status. Once it succeeds, retrieve the result URL. Check the existing task and your usage before retrying so you don't create duplicates.

Need more help?

Submit a ticket to follow up on integration issues.

Technical support ticketSubmit an issue and track its progress