diff --git a/README.md b/README.md
index ddf58e8..c7c4182 100644
--- a/README.md
+++ b/README.md
@@ -2,7 +2,7 @@
 
 A simple common-line interface for [ChatGPT API](https://platform.openai.com/docs/api-reference/chat/create).
 
-- 🌟 Streaming output! 
+- 🌟 Streaming output!
 - 🤖 Both interactive & one-shot mode
 
 <img width="1022" alt="Screen Shot 2023-03-07 at 09 30 53" src="https://user-images.githubusercontent.com/10192522/223295925-00eed881-cdfc-4f46-9510-1e0bd1c99e60.png">
@@ -19,9 +19,11 @@ cargo install --path .
 
 You'll need a OpenAI API key (you can get one [here](https://platform.openai.com/account/api-keys)), and you'll need to export your API Key as an environment variable:
 
+You can also set a OpenAI API base environment variable, just like [openai-python](https://github.com/openai/openai-python/blob/main/openai/__init__.py#L37)
 
 ```
 export OPENAI_API_KEY=<your api key>
+# export OPENAI_API_BASE="https://api.openai.com/v1"
 ```
 
 Then you can start a conversation with ChatGPT:
@@ -33,5 +35,5 @@ heygpt how to record screen on mac
 OR use the interactive mode by providing no prompt:
 
 ```
-heygpt 
+heygpt
 ```
diff --git a/src/main.rs b/src/main.rs
index a3eb76e..f245bf4 100644
--- a/src/main.rs
+++ b/src/main.rs
@@ -192,9 +192,12 @@ async fn complete_and_print(
         format!("Bearer {}", openai_api_key).parse().unwrap(),
     );
 
+    let key = "OPENAI_API_BASE";
+    let openai_api_base = env::var(key).unwrap_or("https://api.openai.com/v1".into());
+
     let client = Client::new();
     let req_builder = client
-        .post("https://api.openai.com/v1/chat/completions".to_string())
+        .post(format!("{}/chat/completions", openai_api_base))
         .headers(headers)
         .json(&data);