Error generating episode - Failed to generate audio: Error code: 404

[atxjd]

I'm using the latest pull and docker-compose profile multi.

When creating a podcast, I'm getting the following error:

Error generating episode - Failed to generate audio: Error code: 404 - {'error': {'code': None, 'message': 'Invalid URL (POST /v1/audio/speech)', 'param': None, 'type': 'invalid_request_error'}}

The output from the console:

open_notebook-1  | 2025-06-14 11:21:30.184 | INFO     | open_notebook.plugins.podcasts:generate_episode:125 - Generating episode  with config {'max_num_chunks': 8, 'min_chunk_size': 600, 'conversation_style': ['Educational'], 'roles_person1': ['Host', 'Conversationalist'], 'roles_person2': ['Expert Guest'], 'dialogue_structure': ['Deep Dive into Subtopics', 'Expert Interviews', 'Key Takeaways'], 'podcast_name': 'TEST Name', 'podcast_tagline': 'Test Tagline, 'output_language': 'English', 'user_instructions': 'Test instructions', 'engagement_techniques': ['Open-ended Questions', 'Evoking Curiosity', 'Spiritual'], 'creativity': 0.2, 'text_to_speech': {'output_directories': {'transcripts': './data/podcasts/transcripts', 'audio': './data/podcasts/audio'}, 'temp_audio_dir': './data/podcasts/audio/tmp', 'ending_message': "Thank you for listening to this episode. Don't forget to subscribe to our podcast for more interesting conversations.", 'default_tts_model': 'openai', 'openai': {'default_voices': {'question': 'alloy', 'answer': 'nova'}, 'model': 'gpt-4o-mini'}, 'audio_format': 'mp3'}} and using model gpt-4o-mini, tts model openai
open_notebook-1  | /app/.venv/lib/python3.12/site-packages/langsmith/client.py:272: LangSmithMissingAPIKeyWarning: API key must be provided when using hosted LangSmith API
open_notebook-1  |   warnings.warn(
open_notebook-1  | Error converting text to speech: Failed to generate audio: Error code: 404 - {'error': {'code': None, 'message': 'Invalid URL (POST /v1/audio/speech)', 'param': None, 'type': 'invalid_request_error'}}
open_notebook-1  | 2025-06-14 11:22:02,596 - podcastfy.client - ERROR - An error occurred in the process_content function: Failed to generate audio: Error code: 404 - {'error': {'code': None, 'message': 'Invalid URL (POST /v1/audio/speech)', 'param': None, 'type': 'invalid_request_error'}}
open_notebook-1  | 2025-06-14 11:22:02,596 - podcastfy.client - ERROR - An error occurred: Failed to generate audio: Error code: 404 - {'error': {'code': None, 'message': 'Invalid URL (POST /v1/audio/speech)', 'param': None, 'type': 'invalid_request_error'}}
open_notebook-1  | 2025-06-14 11:22:02.596 | ERROR    | open_notebook.plugins.podcasts:generate_episode:147 - Failed to generate episode : Failed to generate audio: Error code: 404 - {'error': {'code': None, 'message': 'Invalid URL (POST /v1/audio/speech)', 'param': None, 'type': 'invalid_request_error'}}

docker-compose.yml:

# Instructions on how to use the different compose profiles
# 1. Run `docker compose --profile single up` to start the app and database on the same container
# 2. Run `docker compose --profile multi up` to start the multi container with app and database separate
# 3. Run `docker compose --profile db_only up` to start the database only -- useful if developing locally
services:
  surrealdb:
    image: surrealdb/surrealdb:v2
    ports:
      - "8000:8000"
    volumes:
      - ./surreal_data:/mydata
    environment:
      - SURREAL_EXPERIMENTAL_GRAPHQL=true
    command: start --log info --user root --pass root rocksdb:/mydata/mydatabase.db
    pull_policy: always
    user: root
    profiles: [db_only, multi]
    networks:
      - app-net
  open_notebook:
    image: lfnovo/open_notebook:latest
    ports:
      - "8080:8502"
    env_file:
      - ./docker.env
    depends_on:
      - surrealdb
    pull_policy: always
    profiles: [multi]
    volumes:
      - ./notebook_data:/app/data
    networks:
      - app-net
  open_notebook_single:
    build:
      context: .
      dockerfile: Dockerfile.single
    ports:
      - "8080:8502"
    profiles:
      - single
    volumes:
      - ./.docker_data/data:/app/data
      - ./docker2.env:/app/.env
      - ./google-credentials.json:/app/google-credentials.json

networks:
  app-net:

[lfnovo]

Can you show me your docker.env config? Hide the keys, of course :)

[atxjd]

docker.env

# OPENAI
OPENAI_API_KEY=sk-proj-REDACTED


# ANTHROPIC
# ANTHROPIC_API_KEY=

# GEMINI
# this is the best model for long context and podcast generation
GEMINI_API_KEY=A---REDACTED

# VERTEXAI
# VERTEX_PROJECT=my-google-cloud-project-name
# GOOGLE_APPLICATION_CREDENTIALS=./google-credentials.json
# VERTEX_LOCATION=us-east5

# MISTRAL
# MISTRAL_API_KEY=

# DEEPSEEK
# DEEPSEEK_API_KEY=

# OLLAMA
# OLLAMA_API_BASE="http://10.20.30.20:11434"

# OPEN ROUTER
# OPENROUTER_BASE_URL="https://openrouter.ai/api/v1"
# OPENROUTER_API_KEY=

# GROQ
# GROQ_API_KEY=

# XAI
# XAI_API_KEY=

# ELEVENLABS
# Used only by the podcast feature
# ELEVENLABS_API_KEY=

# VOYAGE AI
# VOYAGE_API_KEY=

# AZURE OPENAI
# AZURE_OPENAI_API_KEY=
# AZURE_OPENAI_ENDPOINT=
# AZURE_OPENAI_API_VERSION="2024-12-01-preview"
# AZURE_OPENAI_DEPLOYMENT_NAME=

# USE THIS IF YOU WANT TO DEBUG THE APP ON LANGSMITH
# LANGCHAIN_TRACING_V2=true
# LANGCHAIN_ENDPOINT="https://api.smith.langchain.com"
# LANGCHAIN_API_KEY=
# LANGCHAIN_PROJECT="Open Notebook"


# CONNECTION DETAILS FOR YOUR SURREAL DB
# Use surrealdb if using docker-compose or add your server ip if using a different setup
SURREAL_ADDRESS="surrealdb"
SURREAL_PORT=8000
SURREAL_USER="root"
SURREAL_PASS="root"
SURREAL_NAMESPACE="open_notebook"
SURREAL_DATABASE="staging"


# FIRECRAWL - Get a key at https://firecrawl.dev/
FIRECRAWL_API_KEY=

# JINA - Get a key at https://jina.ai/
JINA_API_KEY=

[lfnovo]

Can you also show me what's the model your are using for tts and also your podcast configuration?

[atxjd]

I'm not sure what you mean by 'podcast configuration.' Does the JSON in the original message help at all?

[lfnovo]

Here is what I need to see..

Your text to speech models

Your podcast template (Podcast -> Templates)
(show me the whole thing, specially the models and voices part in the end)

[atxjd]

[lfnovo]

OK.. the Audio Model Provider on your podcast config is wrong for openai text to speech..
it should be either tts-1, tts-1-hd or gpt-4o-mini-tts

Can you try it like that?

[atxjd]

Changing to gpt-40-mini-tts worked great. Thank you!