Workflow · Production
Building semantic search and a RAG chat assistant over podcast transcripts with ApertureDB, Cohere, and LangChain
The problem
The author wanted to correlate learnings across podcast episodes, revisit earlier ideas, and share summaries — requiring a queryable, conversational interface over transcript data.
Workflow diagram · grounded in source
1
User submits query
trigger
“I'm guiding you through the art of querying podcast transcripts for quick information retrieval”
2
Embed query with Cohere
ai_action
“the embedding model to embed the user query should be the same as the one you used to embed the data. In this case, I used Cohere's embed-v3”
3
Semantic search in ApertureDB
ai_action
“In ApertureDB, the method "find_similar", which is a Python wrapper for the Descriptors (embeddings) API, enables the user to find similar items (answers in our case) to the provided query”
4
MMR retriever reranks results
ai_action
“I passed search_type to be Maximal Marginal Relevance (MMR) that helps return results with reduced redundancy while maintaining the relevancy between the query and results”
5
LLM generates chat answer
ai_action
“In this case, I used Cohere's Command R model. This model is a strong choice for creating a chat assistant due to its ability to understand the intent behind messages, remember conversation history, and navigate through multi-turn conver…”
6
Answer returned to user
output
“The last step is to invoke the chain where you can pass the question you want to ask the assistant followed by the answer in its raw output format”
Reported outcome
The system enables semantic search and a conversational chat assistant over podcast transcripts, though the author acknowledges limitations in accurately retrieving nuanced information.
Reported stack
ApertureDBWhisperLangChainCommand-R-plusembed-v3
Source
https://mlops.community/blog/semantic-search-to-glean-valuable-insights-from-podcast-part-3
Read source ↗Frequently asked questions
What did this team achieve with this AI workflow?
The system enables semantic search and a conversational chat assistant over podcast transcripts, though the author acknowledges limitations in accurately retrieving nuanced information.
What tools did this team use?
ApertureDB, Whisper, LangChain, Command-R-plus, embed-v3.
How is this workflow AI workflow structured?
User submits query → Embed query with Cohere → Semantic search in ApertureDB → MMR retriever reranks results → LLM generates chat answer → Answer returned to user.