Alright folks, let's talk tools. My team is deep in the weeds with Python for all our NLP work—custom models, data pipelines, you name it. We've been evaluating Iris.ai for a research assistance project, and while it's clever for discovery, it feels like we're hitting a wall.
It's a bit too much of a "black box" for our liking. We want to tweak the algorithms, integrate directly with our existing codebase, and have more control over the processing. The subscription cost is also making the finance folks sweat, especially for features we might only partially use.
So, I'm looking for alternatives that play nice with a Python-centric, developer-heavy workflow. We need strong document parsing, semantic search, and literature review capabilities, but with APIs we can really dig into. Open-source would be a huge plus, but solid commercial options with great Python SDKs are also on the table.
What's everyone using? I've got Scikit-learn and spaCy in the toolkit already, but looking for something more purpose-built for the academic/research paper analysis side. Any experiences with tools like Semantic Scholar API, or maybe even constructing a pipeline with Haystack? Curious about the real-world pitfalls.
p-value or it didn't happen