Two env vars. Same workflow.

If your tool lets you configure an OpenAI base URL and key, it works. Same pattern as Cursor/Cline/aider — for everything else.

1. 1

   The two values to set

   Base URL / API host:  https://api.aicheap.io/v1
   API key:             sk-...
   Model:               gpt-5.4  # or any id from /models

   Copy
2. 2

   Tools known to work this way

   Verified by users in the community: LibreChat, Open WebUI, ChatBox, BoltAI, TypingMind, Jan, BetterChatGPT, Big-AGI, Lobe Chat, NextChat, ChatGPT-Next-Web, AnythingLLM, llm (Simon Willison's CLI). Anything that exposes an "OpenAI base URL" / "OpenAI-compatible endpoint" setting will work — if you find one that doesn't, ping us in the Telegram community.

Anthropic's official CLI agent. Set two env vars and the CLI talks to aicheap.io instead of api.anthropic.com.

1. 1

   Get an API key

   Sign in at console.aicheap.io, pick a plan, and copy the key from the Keys page.

2. 2

   Export the env vars

   export ANTHROPIC_BASE_URL="https://api.aicheap.io/anthropic/v1"
   export ANTHROPIC_API_KEY="sk-..."
   
   # Optional: pin to a specific model
   export ANTHROPIC_MODEL="gpt-5.4"

   Copy
3. 3

   Run claude as usual

   claude
   # or with a prompt
   claude "refactor this function"

   Copy

OpenAI's coding agent. Point OPENAI_BASE_URL at the Codex-compatible endpoint.

1. 1

   Export the env vars

   export OPENAI_BASE_URL="https://api.aicheap.io/openai/v1"
   export OPENAI_API_KEY="sk-..."

   Copy
2. 2

   Run codex

   codex

   Copy

OpenAI-compatible coding agent. Point OPENAI_BASE_URL at our /v1 endpoint.

1. 1

   Export the env vars

   export OPENAI_BASE_URL="https://api.aicheap.io/v1"
   export OPENAI_API_KEY="sk-..."

   Copy
2. 2

   Run opencode

   opencode

   Copy

Add aicheap as a custom provider in your OpenClaw JSON config.

1. 1

   Edit ~/.openclaw/openclaw.json

   // ~/.openclaw/openclaw.json
   {
     "models": {
       "providers": {
         "aicheap": {
           "baseUrl": "https://api.aicheap.io/v1",
           "apiKey": "sk-...",
           "api": "openai-completions"
         }
       }
     }
   }

   Copy
2. 2

   Restart OpenClaw and pick the aicheap provider

   Models served via aicheap appear under the aicheap provider in the model picker.

Override OpenAI in Cursor settings. Works for both chat and tab completion.

1. 1

   Open Cursor → Settings → Models

   Toggle on Override OpenAI Base URL.

2. 2

   Set base URL and key

   Base URL: https://api.aicheap.io/cursor/v1
   OpenAI API Key: sk-...

   Copy
3. 3

   Verify and use

   Click Verify in settings. After it confirms, every Cursor request flows through aicheap.io.

Cline supports any OpenAI-compatible endpoint via its provider settings.

1. 1

   Open the Cline sidebar → ⚙ Settings

   Set API Provider to OpenAI Compatible.

2. 2

   Configure the endpoint

   Base URL: https://api.aicheap.io/v1
   API Key: sk-...
   Model ID: gpt-5.4  # or any model from /models

   Copy

aider speaks OpenAI and Anthropic natively. Both env styles work.

1. 1

   Anthropic-style (Claude models)

   export ANTHROPIC_BASE_URL="https://api.aicheap.io/anthropic/v1"
   export ANTHROPIC_API_KEY="sk-..."
   aider --model gpt-5.4

   Copy
2. 2

   OpenAI-style (GPT, Gemini, Grok, …)

   export OPENAI_API_BASE="https://api.aicheap.io/v1"
   export OPENAI_API_KEY="sk-..."
   aider --model gpt-5.4

   Copy

Add aicheap as an OpenAI-compatible provider in ~/.continue/config.json.

1. 1

   Edit ~/.continue/config.json

   {
     "models": [
       {
         "title": "aicheap — Claude Opus",
         "provider": "openai",
         "model": "gpt-5.4",
         "apiBase": "https://api.aicheap.io/v1",
         "apiKey": "sk-..."
       }
     ]
   }

   Copy

Drop-in replacement. Override base_url when constructing the client.

1. 1

   Python

   from openai import OpenAI
   
   client = OpenAI(
       base_url="https://api.aicheap.io/v1",
       api_key="sk-...",
   )
   
   resp = client.chat.completions.create(
       model="gpt-5.4",
       messages=[{"role": "user", "content": "hello"}],
   )
   print(resp.choices[0].message.content)

   Copy
2. 2

   Node

   import OpenAI from "openai";
   
   const client = new OpenAI({
     baseURL: "https://api.aicheap.io/v1",
     apiKey: "sk-...",
   });
   
   const resp = await client.chat.completions.create({
     model: "gpt-5.4",
     messages: [{ role: "user", content: "hello" }],
   });

   Copy

Anthropic's official SDK works unchanged — set the base URL.

1. 1

   Python

   from anthropic import Anthropic
   
   client = Anthropic(
       base_url="https://api.aicheap.io/anthropic/v1",
       api_key="sk-...",
   )
   
   msg = client.messages.create(
       model="gpt-5.4",
       max_tokens=1024,
       messages=[{"role": "user", "content": "hello"}],
   )

   Copy
2. 2

   Node

   import Anthropic from "@anthropic-ai/sdk";
   
   const client = new Anthropic({
     baseURL: "https://api.aicheap.io/anthropic/v1",
     apiKey: "sk-...",
   });

   Copy

Sanity-check a key in 5 seconds before wiring it into your tools.

1. 1

   OpenAI-compatible request

   curl https://api.aicheap.io/v1/chat/completions \
     -H "Authorization: Bearer sk-..." \
     -H "Content-Type: application/json" \
     -d '{
       "model": "gpt-5.4",
       "messages": [{"role": "user", "content": "ping"}]
     }'

   Copy
2. 2

   Anthropic-compatible request

   curl https://api.aicheap.io/anthropic/v1/messages \
     -H "x-api-key: sk-..." \
     -H "anthropic-version: 2023-06-01" \
     -H "Content-Type: application/json" \
     -d '{
       "model": "gpt-5.4",
       "max_tokens": 1024,
       "messages": [{"role": "user", "content": "ping"}]
     }'

   Copy

Troubleshooting #