OPEN_SOURCE ↗
REDDIT · REDDIT// 13d agoPRODUCT UPDATE
Lorph Adds Deep Research, Citations
Lorph's latest update folds a deeper research backend into its Ollama-backed chat app, so prompts can fan out into search, scraping, and cited synthesis instead of stopping at a single model answer. It also expands document ingestion, with OCR for images plus PDF, DOCX, XLSX, and code/text support.
// ANALYSIS
This is the kind of update that moves an open-source chat app from novelty to utility. The important part is not the UI polish, but the retrieval loop: query expansion, parallel search, deep scraping, and synthesis with citations.
- –Multi-query planning plus DuckDuckGo Lite and YouTube search make it feel closer to a deep-research engine than a normal chatbot.
- –OCR, PDF, DOCX, XLSX, and code/text ingestion let users ground answers in files, not just web pages.
- –Inline citations, extracted images, and embedded videos are the right trust signals for a research-first product.
- –The tradeoff is that it still relies on Ollama cloud access and proxy-based scraping, so self-hosted is a better fit than fully offline.
- –Its new listing in Awesome AI Web Search should help it reach the exact audience that cares about search-centric LLM tooling.
// TAGS
lorphllmsearchchatbotagentopen-sourceself-hosted
DISCOVERED
13d ago
2026-03-29
PUBLISHED
13d ago
2026-03-29
RELEVANCE
8/ 10
AUTHOR
Fantastic-Market-790