GitHub user OpenPj created a discussion: Proposal & Feedback: Native 
Integration with OpenCrawling for High-Throughput RAG & AI Vector Ingestion

Hey StormCrawler Community! 👋

  We are huge fans of Apache StormCrawler and its stream-based, low-latency 
distributed crawling capabilities on Apache Storm.

  As part of the OpenCrawling project a secure, decoupled Java 25 reference 
implementation of the Open Ingestion Standard (OIS) 
https://github.com/opencrawling/open-ingestion-standard for RAG pipelines and 
enterprise vector databases, we are proposing a native integration between 
Apache StormCrawler and OpenCrawling.

  ### 🌟 Why this Integration?

  By connecting StormCrawler with OpenCrawling, users can stream millions of 
web pages at scale, perform document layout analysis and chunking, generate 
vector embeddings (via Ollama, OpenAI, HuggingFace), and index them directly 
into vector stores (pgvector, Qdrant, Milvus, Vespa, OpenSearch), all with 
native ACL security metadata and MCP (Model Context Protocol) support for LLMs.

  ### 🏗️ Proposed Architecture Highlights

  1. OpenCrawlingBolt (`opencrawling-stormcrawler-bolt`): A dedicated Apache 
Storm Bolt for StormCrawler topologies that normalizes crawled web documents 
directly into OIS-compliant binary/REST streams.
  2. `oc-stormcrawler-repository-connector`: A native repository connector in 
OpenCrawling to dynamically deploy, monitor, and manage StormCrawler topologies 
via the Apache Storm Nimbus REST API.
  ──────
  ### 💬 We’d Love Your Feedback!

  We have opened a detailed proposal on our GitHub repository and would love to 
collaborate with the StormCrawler community to get your thoughts on the design, 
configuration ergonomics, and Bolt architecture:

  👉 GitHub Issue: Integration with Apache StormCrawler for High-Throughput 
Distributed Web Crawling & RAG Ingestion 
https://github.com/opencrawling/opencrawling/issues/94

  Whether you have ideas on Bolt design, metrics correlation, topology 
lifecycle management, or are interested in co-contributing, please drop your 
thoughts in the issue or right here!

  Looking forward to building this bridge together! 🚀

GitHub link: https://github.com/apache/stormcrawler/discussions/2073

----
This is an automatically sent email for [email protected].
To unsubscribe, please send an email to: [email protected]

Reply via email to