diff --git a/README.md b/README.md
index de70377..965f882 100644
--- a/README.md
+++ b/README.md
@@ -28,6 +28,22 @@
 4. Build the server (`npm run build`)
 5. Run (`npm start`)
 
+#### Ollama Configuration Guide
+
+- It's recommended if you can run bigger LLM than 14b parameter.
+- You do not need to provide the API KEY
+- Set LLM_PROVIDER to Ollama (It is going to connect to default ollama endpoint)
+- Set LLM_MODELNAME to the model name you can see from Ollama using the command `ollama ls`
+- It is recommended to set TOKEN_PROCESSING_CHARACTER_LIMIT between 10000-20000 (Approx 300-600 lines of code) if you are using low param LLM (ex. 8b, 14b)
+
+**Example:**
+
+```
+LLM_PROVIDER=ollama
+LLM_APIKEY=
+LLM_MODELNAME=qwen2.5:14b
+```
+
 ### Additional Information
 
 > [!CAUTION]
diff --git a/src/llm/provider/ollama.ts b/src/llm/provider/ollama.ts
new file mode 100644
index 0000000..c80b084
--- /dev/null
+++ b/src/llm/provider/ollama.ts
@@ -0,0 +1,26 @@
+import { ChatOllama } from '@langchain/ollama';
+import LLMConfig from '../llm-config';
+import { HistoryItem, LLMProvider } from '../llm-provider';
+
+export class OllamaProvider extends LLMProvider {
+  private llm: ChatOllama;
+
+  constructor(
+          modelName: string,
+          llmconfig: LLMConfig) {
+    super('', modelName, llmconfig, '');
+    this.llm = new ChatOllama({
+        model: modelName,
+        temperature: llmconfig.temperature,
+        // format: 'json', // Forcing JSON output degraded the quality of the responses
+        topP: llmconfig.topP,
+        topK: llmconfig.topK,
+    });
+  }
+
+  async run(userPrompt: string, history: HistoryItem[]): Promise<string> {
+    const response = await this.llm.invoke(userPrompt);
+    console.log(response.content.toString());
+    return response.content.toString();
+  }
+}
\ No newline at end of file
diff --git a/src/service/llm-factory.ts b/src/service/llm-factory.ts
index e04bd81..adeb323 100644
--- a/src/service/llm-factory.ts
+++ b/src/service/llm-factory.ts
@@ -3,6 +3,7 @@ import DeepSeekProvider from '@/llm/provider/deepseek'
 import GoogleProvider from '@/llm/provider/google'
 import LLMConfig from '@/llm/llm-config'
 import dotenv from 'dotenv'
+import { OllamaProvider } from '@/llm/provider/ollama'
 dotenv.config()
 
 /**
@@ -19,11 +20,21 @@ export default class LLMFactory {
         const apiKey = process.env.LLM_APIKEY
         const modelName = process.env.LLM_MODELNAME
 
+        if (!provider) {
+            throw new Error('LLM Provider is not specified. Please set LLM_PROVIDER in the environment\nExample: LLM_PROVIDER=google, LLM_PROVIDER=deepseek, LLM_PROVIDER=ollama')
+        }
+
+        if(!modelName) {
+            throw new Error('LLM Model name is not specified. Example: LLM_MODELNAME=llama3.3 for llama3.3')
+        }
+
         switch (provider) {
             case 'google':
                 return new GoogleProvider(apiKey!, modelName!, llmConfig)
             case 'deepseek':
                 return new DeepSeekProvider(apiKey!, modelName!, llmConfig)
+            case 'ollama':
+                return new OllamaProvider(modelName!, llmConfig)
             default:
                 throw new Error(`Unsupported LLM provider: ${provider}`)
         }