Book a demo, and we’ll get your data set up up to the free plan. No credit card needed.
Absolutely! Our versatile product is designed to accommodate businesses of all sizes, ranging from startups and small enterprises to large corporations. Whether you’re just starting or already established, our solution offers the scalability and adaptability to meet your unique marketing needs and drive remarkable results.
Yep! We provide dedicated support for new users to ensure a smooth onboarding experience. Feel free to schedule a call with us to get set up with your knowledge base here: https://calendly.com/libraria-ai
No coding required! Our platform offers a user-friendly interface, making it easy for you to create and manage campaigns without any coding knowledge.
Steer conversations with Rich Responses: Customize and control your responses with Libraria’s tools. We make it easy for you to guarantee responses that are on brand, accurate, and makes sense for your users. For each message you send, we make up 2-6 calls to Chat GPT to get the best response for your users.
Internal Linking: Libraria gives you the ability to return links AND additionally link a “Learn More” snippet to each of your responses for users to continue their journey on your site.
Anti-Hallucination tools: Libraria created several “fail safes” to prevent hallucinations. You can predefine a set of questions and answers that will be used verbatim when a customer asks a particular question, or set it to be the only document used to send to Chat GPT for a response. We also have a set of prompts in our documentation to help you refine your prompts to prevent hallucinations. Finally, our “Learn More” snippets will never hallucinate links.
Ingest almost any type of data: Libraria lets you add youtube videos, PDFs, google drive folders, files, docs, docx, pdfs, raw text, urls, xmls OR crawl a website. We accept more information than any other chatbot available.
Preprocessing of information uploaded: Most chatbots only ingest the data without taking into account the TYPE of data you give them. Youtubee links given we break down into timestamps. Source snippets will show which timestamp to go to for each question answered. PDFs give a direct link to the page. Websites send you to the correct page.
Images and Image Snippets: Add images in your responses to make your responses more engaging. Image snippets in conjunction with Learn More snippets are guaranteed to show up when a relevant document is matched.
Detailed Query history: See which documents matched for each query, and which documents were used to generate a response. You can see the “match similarity” rate so that you can make fine tuned adjustments to your knowledge base.
Chatbot, Widget, Landing Page: We provide an “Assistant Widget”, which is more of a Question and Answer tool, a chatbot integration to embed on your website, and a shareable landing page both in assistant and chat view. Libraria makes it as easy as possible for users to ask questions and get answers.
Gating your chatbot: Libraria gives you the option to have your chatbot public, unlisted, protected by a password, and private to only members of your team.
Team features: Libraria gives you the ability to add members to your team to collaborate on your knowledge base. You can also create multiple knowledge bases for different teams.
Automatic Indexing with Pre-filled Questions: When your chatbot is asked what it knows about / has information on, it will give you an up-to-date answer depending on the documents you upload, without having to manually update it to tell it what it knows.
Yes, Libraria lets you create a “custom answer” that will train your AI to respond to a specific question. You can even add images to this too.
Absolutely, Libraria lets you show images via a prompt or via a photo snippet. Check out our documentation on how to add images with https://docs.libraria.ai.
Yes! You can add members to your team to collaborate on your knowledge base. You can also create multiple knowledge bases for different teams.
Upload them all to a google drive folder, and we will recursively scrape through every single PDF.
By default, we use GPT 3.5. Startup and Enterprise plans have the ability to use GPT-4, and Team, Startup and Enterprise plans can run GPT 3.5 with 16K context.