RAG over documents
Chunk, embed, retrieve (vector / hybrid / filtered / conversational), and ground an
answer. Python, Open in Colab.
Agent memory
Long-term memory for an agent: hybrid recall, SQL over memory, and mutation, over
one table. Node.js.
Hybrid search service
An embedded BM25 + vector search API over a product catalog, with no separate search
server. Node.js.
Code search
Exact, natural-language (vector), and keyword (BM25) search over a corpus of
functions. Python, Open in Colab.
Analytics + search
SQL
GROUP BY time-series and leaderboards alongside BM25, over one table. Python,
Open in Colab.Live / mutable data
Keep a live inventory current with
update / delete, durable across reconnect.
Python, Open in Colab.LangChain RAG & agents
Live-data RAG, a research assistant, and a support-ops agent on LangChain.
Python, Open in Colab.
CrewAI crews
A support-triage crew, a content-research crew, and a knowledge crew on CrewAI.
Python, Open in Colab.