diff --git a/assets/box-icon.png b/assets/box-icon.png
new file mode 100644
index 0000000..50f7470
Binary files /dev/null and b/assets/box-icon.png differ
diff --git a/package.json b/package.json
index 95b181f..84b530b 100755
--- a/package.json
+++ b/package.json
@@ -176,12 +176,19 @@
       "description": "Check for extension updates from the official GitHub source",
       "icon": "update-icon.png",
       "mode": "view"
+    },
+    {
+      "name": "configureG4FLocalApi",
+      "title": "Configure GPT4Free Local API",
+      "description": "Configure GPT4Free Local API settings",
+      "icon": "box-icon.png",
+      "mode": "view"
     }
   ],
   "preferences": [
     {
       "name": "gptProvider",
-      "title": "Default GPT Provider",
+      "title": "Default Provider",
       "description": "The default provider and model used in this extension.",
       "required": false,
       "type": "dropdown",
@@ -241,6 +248,10 @@
         {
           "title": "Google Gemini (requires API Key)",
           "value": "GoogleGemini"
+        },
+        {
+          "title": "GPT4Free Local API",
+          "value": "G4FLocal"
         }
       ],
       "default": "GPT35"
diff --git a/src/aiChat.jsx b/src/aiChat.jsx
index d4fb314..c700fe2 100755
--- a/src/aiChat.jsx
+++ b/src/aiChat.jsx
@@ -285,7 +285,7 @@ export default function Chat({ launchContext }) {
         }
       >
         <Form.TextArea id="chatText" title="Chat Transcript" />
-        <Form.Description title="GPT Model" text="The provider and model used for this chat." />
+        <Form.Description title="Provider" text="The provider and model used for this chat." />
         <Form.Dropdown id="provider" defaultValue={providers.default_provider_string()}>
           {providers.ChatProvidersReact}
         </Form.Dropdown>
@@ -416,7 +416,7 @@ export default function Chat({ launchContext }) {
             second: "2-digit",
           })}`}
         />
-        <Form.Description title="GPT Model" text="The provider and model used for this chat." />
+        <Form.Description title="Provider" text="The provider and model used for this chat." />
         <Form.Dropdown id="provider" defaultValue={defaultProviderString}>
           {providers.ChatProvidersReact}
         </Form.Dropdown>
diff --git a/src/api/Providers/deepinfra.jsx b/src/api/Providers/deepinfra.jsx
index e67fa05..6a2e52e 100644
--- a/src/api/Providers/deepinfra.jsx
+++ b/src/api/Providers/deepinfra.jsx
@@ -49,19 +49,17 @@ export const getDeepInfraResponse = async function* (chat, options, max_retries
     const reader = response.body;
     for await (let chunk of reader) {
       const str = chunk.toString();
-
       let lines = str.split("\n");
+
       for (let i = 0; i < lines.length; i++) {
         let line = lines[i];
         if (line.startsWith("data: ")) {
           let chunk = line.substring(6);
           if (chunk.trim() === "[DONE]") return; // trim() is important
 
-          let data = JSON.parse(chunk);
-          let choice = data["choices"][0];
-          // python: if "content" in choice["delta"] and choice["delta"]["content"]:
-          if ("delta" in choice && "content" in choice["delta"] && choice["delta"]["content"]) {
-            let delta = choice["delta"]["content"];
+          try {
+            let data = JSON.parse(chunk);
+            let delta = data["choices"][0]["delta"]["content"];
             if (first) {
               delta = delta.trimStart();
             }
@@ -69,9 +67,9 @@ export const getDeepInfraResponse = async function* (chat, options, max_retries
               first = false;
               yield delta;
             }
-          }
+          } catch (e) {} // eslint-disable-line
         }
-      } // readline
+      }
     }
   } catch (e) {
     if (max_retries > 0) {
diff --git a/src/api/Providers/g4f.jsx b/src/api/Providers/g4f.jsx
index 099dfa5..b0b208a 100644
--- a/src/api/Providers/g4f.jsx
+++ b/src/api/Providers/g4f.jsx
@@ -1,5 +1,6 @@
 import { G4F } from "g4f";
 const g4f = new G4F();
+import { messages_to_json } from "../../classes/message";
 
 export const G4FProvider = {
   GPT: g4f.providers.GPT,
@@ -10,5 +11,8 @@ export const getG4FResponse = async (chat, options) => {
   // replace "g4f_provider" key in options with "provider"
   options.provider = options.g4f_provider; // this is a member of G4FProvider enum
   delete options.g4f_provider;
+
+  chat = messages_to_json(chat);
+
   return await g4f.chatCompletion(chat, options);
 };
diff --git a/src/api/Providers/g4f_local.jsx b/src/api/Providers/g4f_local.jsx
new file mode 100644
index 0000000..dfa013a
--- /dev/null
+++ b/src/api/Providers/g4f_local.jsx
@@ -0,0 +1,136 @@
+// This module allows communication and requests to the local G4F API.
+// Read more here: https://github.com/xtekky/gpt4free/blob/main/docs/interference.md
+
+import { exec } from "child_process";
+import fetch from "node-fetch";
+
+import { Storage } from "../storage";
+import { messages_to_json } from "../../classes/message";
+
+import { environment, Form } from "@raycast/api";
+
+// constants
+const DEFAULT_MODEL = "meta-ai";
+export const DEFAULT_TIMEOUT = "900";
+
+const BASE_URL = "http://localhost:1337/v1";
+const API_URL = "http://localhost:1337/v1/chat/completions";
+const MODELS_URL = "http://localhost:1337/v1/models";
+
+// main function
+export const G4FLocalProvider = "G4FLocalProvider";
+export const getG4FLocalResponse = async function* (chat, options) {
+  if (!(await isG4FRunning())) {
+    await startG4F();
+  }
+
+  chat = messages_to_json(chat);
+  const model = await getSelectedG4FModel();
+
+  const response = await fetch(API_URL, {
+    method: "POST",
+    headers: {
+      "Content-Type": "application/json",
+    },
+    body: JSON.stringify({
+      model: model,
+      stream: options.stream,
+      messages: chat,
+    }),
+  });
+
+  // Important! we assume that the response is a stream, as this is true for most G4F models.
+  // If in the future this is not the case, we should add separate handling for non-streaming responses.
+  const reader = response.body;
+  for await (let chunk of reader) {
+    const str = chunk.toString();
+    let lines = str.split("\n");
+    for (let i = 0; i < lines.length; i++) {
+      let line = lines[i];
+      if (line.startsWith("data: ")) {
+        let chunk = line.substring(6);
+        if (chunk.trim() === "[DONE]") return; // trim() is important
+
+        try {
+          let data = JSON.parse(chunk);
+          let delta = data["choices"][0]["delta"]["content"];
+          if (delta) {
+            yield delta;
+          }
+        } catch (e) {} // eslint-disable-line
+      }
+    }
+  }
+};
+
+/// utilities
+
+// check if the G4F API is running
+// with a request timeout of 0.5 seconds (since it's localhost)
+const isG4FRunning = async () => {
+  try {
+    const controller = new AbortController();
+    setTimeout(() => controller.abort(), 500);
+    const response = await fetch(BASE_URL, { signal: controller.signal });
+    return response.ok;
+  } catch (e) {
+    return false;
+  }
+};
+
+// get available models
+const getG4FModels = async () => {
+  try {
+    const response = await fetch(MODELS_URL);
+    return (await response.json()).data || [];
+  } catch (e) {
+    return [];
+  }
+};
+
+// get available models as dropdown component
+export const getG4FModelsDropdown = async () => {
+  const models = await getG4FModels();
+  return (
+    <Form.Dropdown id="model" title="Model" defaultValue={await getSelectedG4FModel()}>
+      {models.map((model) => {
+        return <Form.Dropdown.Item title={model.id} key={model.id} value={model.id} />;
+      })}
+    </Form.Dropdown>
+  );
+};
+
+// get G4F executable path from storage
+export const getG4FExecutablePath = async () => {
+  return await Storage.read("g4f_executable", "g4f");
+};
+
+// get the currently selected G4F model from storage
+const getSelectedG4FModel = async () => {
+  return await Storage.read("g4f_model", DEFAULT_MODEL);
+};
+
+// get G4F API timeout (in seconds) from storage
+export const getG4FTimeout = async () => {
+  return parseInt(await Storage.read("g4f_timeout", DEFAULT_TIMEOUT)) || parseInt(DEFAULT_TIMEOUT);
+};
+
+// start the G4F API
+const startG4F = async () => {
+  const exe = await getG4FExecutablePath();
+  const timeout_s = await getG4FTimeout();
+  const START_COMMAND = `export PATH="/opt/homebrew/bin:$PATH"; ( ${exe} api ) & sleep ${timeout_s} ; kill $!`;
+  const dirPath = environment.supportPath;
+  try {
+    const child = exec(START_COMMAND, { cwd: dirPath });
+    console.log("G4F API Process ID:", child.pid);
+    child.stderr.on("data", (data) => {
+      console.log("g4f >", data);
+    });
+    // sleep for some time to allow the API to start
+    await new Promise((resolve) => setTimeout(resolve, 2000));
+    console.log(`G4F API started with timeout ${timeout_s}`);
+  } catch (e) {
+    console.log(e);
+  }
+};
diff --git a/src/api/Providers/google_gemini.jsx b/src/api/Providers/google_gemini.jsx
index 40f3fea..9268bed 100644
--- a/src/api/Providers/google_gemini.jsx
+++ b/src/api/Providers/google_gemini.jsx
@@ -43,9 +43,7 @@ export const getGoogleGeminiResponse = async (chat, options, stream_update, max_
           response = await geminiChat.ask(query, { safetySettings: safetySettings });
           return response;
         }
-      } catch (e) {
-        continue;
-      }
+      } catch (e) {} // eslint-disable-line
     }
   } catch (e) {
     // if all API keys fail, we allow a few retries
diff --git a/src/api/gpt.jsx b/src/api/gpt.jsx
index 7ee3b53..b376bc9 100755
--- a/src/api/gpt.jsx
+++ b/src/api/gpt.jsx
@@ -426,6 +426,9 @@ export const chatCompletion = async (chat, options, stream_update = null, status
   } else if (provider === providers.G4FProvider) {
     // G4F
     response = await providers.getG4FResponse(chat, options);
+  } else if (provider === providers.G4FLocalProvider) {
+    // G4F Local
+    response = await providers.getG4FLocalResponse(chat, options);
   }
 
   // stream = false
diff --git a/src/api/providers.jsx b/src/api/providers.jsx
index 691d76b..1d1dfd4 100644
--- a/src/api/providers.jsx
+++ b/src/api/providers.jsx
@@ -33,8 +33,12 @@ export { ReplicateProvider, getReplicateResponse };
 import { GeminiProvider, getGoogleGeminiResponse } from "./Providers/google_gemini";
 export { GeminiProvider, getGoogleGeminiResponse };
 
+// G4F Local module
+import { G4FLocalProvider, getG4FLocalResponse } from "./Providers/g4f_local";
+export { G4FLocalProvider, getG4FLocalResponse };
+
 /// All providers info
-// {provider, model, stream, extra options}
+// { provider internal name, {provider, model, stream, extra options} }
 // prettier-ignore
 export const providers_info = {
   GPT35: { provider: NexraProvider, model: "chatgpt", stream: true },
@@ -51,6 +55,7 @@ export const providers_info = {
   ReplicateLlama3_70B: { provider: ReplicateProvider, model: "meta/meta-llama-3-70b-instruct", stream: true },
   ReplicateMixtral_8x7B: { provider: ReplicateProvider, model: "mistralai/mixtral-8x7b-instruct-v0.1", stream: true },
   GoogleGemini: { provider: GeminiProvider, model: "gemini-1.5-flash-latest", stream: true },
+  G4FLocal: { provider: G4FLocalProvider, stream: true },
 };
 
 /// Chat providers (user-friendly names)
@@ -69,6 +74,7 @@ export const chat_providers = [
   ["Replicate (meta-llama-3-70b)", "ReplicateLlama3_70B"],
   ["Replicate (mixtral-8x7b)", "ReplicateMixtral_8x7B"],
   ["Google Gemini (requires API Key)", "GoogleGemini"],
+  ["GPT4Free Local API", "G4FLocal"],
 ];
 
 export const ChatProvidersReact = chat_providers.map((x) => {
diff --git a/src/api/storage.jsx b/src/api/storage.jsx
index 50c76b8..a50d09a 100644
--- a/src/api/storage.jsx
+++ b/src/api/storage.jsx
@@ -109,7 +109,8 @@ export const Storage = {
     await Storage.run_sync();
   },
 
-  // combined read function - read from local storage, fallback to file storage
+  // combined read function - read from local storage, fallback to file storage.
+  // also writes the default value to local storage if it is provided and key is not found
   read: async (key, default_value = undefined) => {
     let value;
     if (await Storage.localStorage_has(key)) {
diff --git a/src/configureG4FLocalApi.jsx b/src/configureG4FLocalApi.jsx
new file mode 100644
index 0000000..942d090
--- /dev/null
+++ b/src/configureG4FLocalApi.jsx
@@ -0,0 +1,64 @@
+import { getG4FExecutablePath, getG4FTimeout, DEFAULT_TIMEOUT, getG4FModelsDropdown } from "./api/Providers/g4f_local";
+import { Storage } from "./api/storage";
+import { help_action } from "./helpers/helpPage";
+
+import { Form, ActionPanel, Action, useNavigation, showToast, Toast } from "@raycast/api";
+import { useState, useEffect } from "react";
+
+export default function ConfigureG4FLocalApi() {
+  const [executablePath, setExecutablePath] = useState("");
+  const [timeout, setTimeout] = useState("");
+  const [modelsDropdown, setModelsDropdown] = useState([]);
+  const [rendered, setRendered] = useState(false);
+
+  const { pop } = useNavigation();
+
+  useEffect(() => {
+    (async () => {
+      setExecutablePath(await getG4FExecutablePath());
+      setTimeout((await getG4FTimeout()).toString());
+      setModelsDropdown(await getG4FModelsDropdown());
+      setRendered(true);
+    })();
+  }, []);
+
+  return (
+    <Form
+      actions={
+        <ActionPanel>
+          <Action.SubmitForm
+            title="Save"
+            onSubmit={async (values) => {
+              pop();
+              await Storage.write("g4f_executable", values.g4f_executable);
+              await Storage.write("g4f_timeout", values.g4f_timeout || DEFAULT_TIMEOUT);
+              await Storage.write("g4f_model", values.model);
+              await showToast(Toast.Style.Success, "Configuration Saved");
+            }}
+          />
+          {help_action("g4fLocal")}
+        </ActionPanel>
+      }
+    >
+      <Form.Description text="Configure the GPT4Free Local API. Select 'Help' for the full guide." />
+      <Form.TextField
+        id="g4f_executable"
+        title="G4F Executable Path"
+        value={executablePath}
+        onChange={(x) => {
+          if (rendered) setExecutablePath(x);
+        }}
+      />
+      <Form.TextField
+        id="g4f_timeout"
+        title="G4F API Timeout (in seconds)"
+        info="After this timeout, the G4F API will be stopped. It will be automatically started again when used. This saves resources when the API is not in use."
+        value={timeout}
+        onChange={(x) => {
+          if (rendered) setTimeout(x);
+        }}
+      />
+      {modelsDropdown}
+    </Form>
+  );
+}
diff --git a/src/genImage.jsx b/src/genImage.jsx
index 46b6823..2dca514 100644
--- a/src/genImage.jsx
+++ b/src/genImage.jsx
@@ -120,7 +120,7 @@ export default function genImage() {
             if (err) {
               toast(Toast.Style.Failure, "Error saving image");
               console.log("Error saving image. Current path: " + __dirname);
-              console.error(err);
+              console.log(err);
               imagePath = "";
             }
           });
diff --git a/src/helpers/helpPage.jsx b/src/helpers/helpPage.jsx
index 0db5991..808ce94 100644
--- a/src/helpers/helpPage.jsx
+++ b/src/helpers/helpPage.jsx
@@ -5,6 +5,7 @@ const helpPages = {
   aiChat: "https://github.com/XInTheDark/raycast-g4f/wiki/Help-page:-AI-Chat",
   customAICommands: "https://github.com/XInTheDark/raycast-g4f/wiki/Help-page:-Custom-AI-Commands",
   genImage: "https://github.com/XInTheDark/raycast-g4f/wiki/Help-page:-Generate-Images",
+  g4fLocal: "https://github.com/XInTheDark/raycast-g4f/wiki/Help-page:-GPT4Free-Local-API",
   default: "https://github.com/XInTheDark/raycast-g4f/blob/main/README.md",
 };