OraClaw -- deterministic decision tools for RAG pipelines (scoring, ranking, anomaly detection) #13861
Whatsonyourmind
started this conversation in
Show and tell
OraClaw -- deterministic decision tools for RAG pipelines (scoring, ranking, anomaly detection)
#13861
Replies: 0 comments
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
-
Hey RAGFlow community! Sharing tools that can enhance RAG pipeline decision-making.
The idea
RAG pipelines often need to make structured decisions: Which retrieval results to rank highest? How to score document relevance? When to flag anomalous queries? These are mathematical problems that LLMs solve expensively and inconsistently.
OraClaw provides 19 deterministic algorithms that handle these decisions in sub-25ms.
Most relevant for RAG workflows
Available as
https://oraclaw-api.onrender.comnpx @oraclaw/mcp-server@oraclaw/graph,@oraclaw/bandit,@oraclaw/detect, etc.MIT licensed, 1,072 tests. Would love to hear thoughts on integrating decision algorithms into RAG pipelines.
Links: GitHub | npm | Demo
Beta Was this translation helpful? Give feedback.
All reactions