Viktor.comAn AI coworker that actually does the work
Docubix
Turn your docs into an AI assistant that cites its sources
7 followers
Turn your docs into an AI assistant that cites its sources
7 followers
Upload PDFs and documents to create AI assistants that answer questions using your own content-with cited sources. Deploy through our dashboard or REST API in minutes, without building or maintaining a RAG pipeline.
Hi everyone! I'm Toheed, the solo founder behind Docubix.
Every company already has the answers - they're just scattered across PDFs, docs, help centers, and internal files. Finding them shouldn't require weeks of RAG infrastructure work.
Docubix lets you upload your existing content and get an AI assistant that answers questions from it - with cited sources - through our dashboard or a REST API. No pipeline to build, no models to fine-tune.
It's built first for customer support and internal knowledge teams, and works well for SaaS docs and course/coaching content too.
I'd love your feedback:
- Where does the upload → ask-a-question flow feel unclear?
- What would make you trust an AI-generated answer enough to show it to your customers?
- Would you reach for an API integration first, or an embeddable widget?
I'll be here all day to answer everything. Thanks for checking out Docubix.
How does the pricing scale with the number of documents or queries, and is there a cap on file size or pages per upload?
@cemilqlcd Thanks for the question. We currently offer a Free plan and a Pro plan. The Free plan includes 100 queries/month, 1 knowledge base, 20 documents, and 250 MB of storage. The Pro plan includes 3,000 queries/month, 10 knowledge bases, 1,000 documents, and 10 GB of storage.
Individual uploads are currently limited to 50 MB per file. We don't enforce a page limit-the practical limit is the file size. If you have larger knowledge bases or enterprise-scale requirements, we'd love to learn more about your use case and can work with you on higher limits as Docubix evolves.
The cited sources in the answer view look really well integrated, the inline references feel native rather than tacked on. Nice execution on something a lot of RAG tools fumble.
@betlkonurabzyh Thanks, Betül! That means a lot. Citations were one of the features we cared about most-we wanted answers to be easy to verify, not just accurate. Glad it stood out.