@@ -11,6 +11,7 @@ AI-powered application that transforms PDF documents into engaging podcast-style
1111- [ Quick Start Deployment] ( #quick-start-deployment )
1212- [ User Interface] ( #user-interface )
1313- [ Troubleshooting] ( #troubleshooting )
14+ - [ Additional Info] ( #additional-info )
1415
1516---
1617
@@ -139,6 +140,22 @@ OpenAI API Key for text-to-speech:
139140- Key format starts with ` sk-proj- `
140141- Requires access to TTS APIs
141142
143+ ### Local Development Configuration
144+
145+ ** For Local Testing Only (Optional)**
146+
147+ If you're testing with a local inference endpoint using a custom domain (e.g., ` inference.example.com ` mapped to localhost in your hosts file):
148+
149+ 1 . Edit ` api/llm-service/.env ` and set:
150+ ``` bash
151+ LOCAL_URL_ENDPOINT=inference.example.com
152+ ```
153+ (Use the domain name from your INFERENCE_API_ENDPOINT without ` https:// ` )
154+
155+ 2 . This allows Docker containers to resolve your local domain correctly.
156+
157+ ** Note:** For public domains or cloud-hosted endpoints, leave the default value ` not-needed ` .
158+
142159---
143160
144161## Quick Start Deployment
@@ -311,3 +328,12 @@ docker compose down -v
311328For detailed troubleshooting guidance and solutions to common issues, refer to:
312329
313330[ TROUBLESHOOTING_and_ADDITIONAL_INFO.md] ( ./TROUBLESHOOTING_and_ADDITIONAL_INFO.md )
331+
332+ ## Additional Info
333+
334+ The following models have been validated with pdf-podcast:
335+
336+ | Model | Hardware |
337+ | -------| ----------|
338+ | ** deepseek-ai/DeepSeek-R1-Distill-Qwen-32B** | Gaudi |
339+ | ** Qwen/Qwen3-4B-Instruct-2507** | Xeon |
0 commit comments