Tool Integrations
Detailed setup notes for Cursor, Codex, Cline, Claude Code, and OpenCode.
Quick Start
Use this page when you want a working request first. It focuses on the minimum setup for curl, Codex, and Claude Code, then points you to the next pages for deeper configuration.
Before You Start
Prepare one API key, one model group, and one model name before testing anything else.
Base URL for OpenAI-compatible requests: https://api.lingcorex.ai/v1
Base URL for Anthropic-native requests: https://api.lingcorex.ai
Create the API key in the dashboard first, then pick the group that matches your tool or endpoint format.
If your tool expects OpenAI-compatible calls, use a group intended for that path. If it expects Anthropic-native calls, use an Anthropic-native group instead.
If you are unsure which group to use, read API Keys & Groups next. That page explains the group concept and how to identify Anthropic endpoint groups.
curl
The fastest connectivity check is a single OpenAI-compatible chat completions request.
OpenAI-compatible curl request
curl https://api.lingcorex.ai/v1/chat/completions \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "gpt-5.4-mini",
"messages": [
{"role": "user", "content": "Return a one-line health check."}
],
"stream": false
}'A 401 usually means the API key is missing or invalid. A 402 usually means the billing account does not have enough balance for the estimated pre-authorization.
Cursor
If you use Cursor, follow this real-screen walkthrough to set the API key, override the OpenAI-compatible Base URL, choose a supported official model, and verify the first chat message.
Codex
Codex uses the OpenAI-compatible route, so the minimum setup is just a base URL and an API key.
01
Create an OpenAI-compatible key
In the dashboard, create an API key from a group that is meant for OpenAI-compatible tools.
02
Set the environment variables
Export OPENAI_BASE_URL=https://api.lingcorex.ai/v1 and OPENAI_API_KEY=<YOUR_API_KEY> before launching Codex.
03
Confirm the first request
Run a short Codex session against a lightweight model first, then move into project-specific config if you want a persistent setup.
Claude Code
Claude Code works best when the key comes from an Anthropic-native group, because that path preserves Anthropic-native behavior more faithfully.
Codex shell setup
export OPENAI_BASE_URL=https://api.lingcorex.ai/v1
export OPENAI_API_KEY=YOUR_API_KEY
codexClaude Code shell setup
ANTHROPIC_BASE_URL=https://api.lingcorex.ai \
ANTHROPIC_API_KEY=YOUR_API_KEY \
ANTHROPIC_CUSTOM_MODEL_OPTION=claude-sonnet-4-5 \
claudeUse Claude Code when you want Anthropic-native request behavior, especially if caching controls matter to your workflow.
Next Steps
Once the first request works, use the pages below to make the integration production-ready.
Caching
How cache behavior changes across interfaces, and where LingCoreX optimizes cache usage automatically.
API Keys & Groups
How to create keys, set QPM, set daily spend limits, and choose the right group for the endpoint you need.