Aginsi vs Humata

Aginsi is an AI research co-pilot that turns reading material into summaries with key passages highlighted. It accepts articles, academic papers, PDFs, long threads, reports, and video transcripts; summaries keep their references and show the amount of compression and estimated reading time saved. The free Margin plan costs nothing and includes 10 monthly summaries for articles and PDFs. Scholar costs $9 a month and adds unlimited summaries, a library with tags and search, Markdown and PDF export, and a browser companion. Faculty costs $29 a month and adds team seats, shared shelves, batch and API access, and priority support. The product is available through web, extension, and API. Aginsi says it discards submitted source documents and retains only summaries users choose to save. Source text is processed transiently and not used to train models. Users can request access to, correction of, export of, or deletion of account data and saved summaries by contacting [email protected]. The company describes itself as a small studio in Helsinki.

Humata is a document question-and-answer service for finding information across uploaded files. Ask questions about document contents, search for answers, summarize findings, or compare documents; responses include citations that link to relevant sections in source files. Users can also embed Humata’s PDF AI in a webpage, while its API provides programmatic access to documents, conversations, and answers. The product is available on the web and through an API, with file uploads supported. The Free plan costs nothing and includes one user and up to 60 pages of monthly use. Expert costs $999 a month for three users and 500 pages, with $0.02 per additional page. Team costs $49 a month per user for 10 users and 5,000 pages, with $0.01 per additional page. Humata supports SAML 2.0 sign-on and states that it encrypts customer data at rest and in transit, is SOC 2 Type II compliant, and does not use user-provided data to train its models.