Powerful AI for Your Documents
Whether It’s Searching Your Files or the Web, Responses Are Richly Cited (File Name & Page Number or URL), and Grounded in Actual Source Texts, which are Readily Available for you to view.
Reliable Results
Advanced Document Retrival
Large Language Models are very powerful, but have major drawbacks. Leveraging advanced RAG techniques, we improve the model's output by enhancing accuracy, contextual relevance, and reliability in the following ways:
- Credibility and Transparency
For every piece of information, we provide a reference to its source document and page, allowing for independent verification, significantly boosting credibility and reliability. This transparent referencing also sheds light on the AI’s reasoning, offering users a clear understanding of how the responses are derived.
- Overcoming Memory Limitations
Traditional language models have a "token limit", which restricts large text processing. Our use of vector embeddings compresses text data, enabling efficient handling of extensive document collections and allows users to query hundreds of documents simultaneously, overcoming standard model limitations.
- Contextual Relevance
By constraining the model's response to the context of the referenced documents and maintaining strict relevance to the source material, we significantly reduce the occurance of "hallucinations" – instances where an AI might generate a plausible but incorrect response.
Pricing
Simple pricing for everyone
Unused message credits rollover automatically.
Starter
A plan to streamline your productivity.
$9.99/month
Buy plan- Ability to chat with multiple documents at once
- 10,000 page upload limit, you may delete uploaded documents to restore page credits
- 300 message credits per month. Any unused message credits rollover automatically.
Pro
A plan for the productivity power users.
$14.99/month
Buy plan- Ability to chat with multiple documents at once
- 22,500 page upload limit, you may delete uploaded documents to restore page credits
- 600 message credits per month. Any unused message credits rollover automatically.
Frequently asked questions
Have a different question and can’t find the answer you’re looking for? Reach out to our support team by sending us an email and we’ll get back to you as soon as we can.
- What exactly is RAG and how does it enhance information retrieval?
- RAG, short for Retrieval-Augmented Generation, is an AI technique that first fetches relevant information based on your query and then crafts responses that are informed by this data. This process ensures that the answers you get are not only precise but also deeply rooted in the specific content you're querying.
- What if my documents contain images?
- Our system is designed for text analysis; therefore, any images within documents will be ignored during processing. For those looking to include text from images in their queries, we recommend using an OCR (Optical Character Recognition) tool to convert the images to text before uploading the documents.
- How does your system handle my document data?
- We take privacy seriously, instead of storing actual documents, we convert the text into numerical embeddings and securely store these, along with some associated metadata, in Pinecone —a dedicated, secure vector database designed for efficiency and confidentiality.
- What exactly are embeddings?
- Embeddings are a form of data representation where text is converted into a vector of numbers, enabling the machine to understand and process language. These vectors capture the semantic meaning of words or sentences, allowing for complex operations like searching, similarity comparisons, and machine learning tasks on textual data.
- What document types do you support?
- Our RAG system supports a variety of document types including PDFs, Word documents, and plain text files, enabling you to query a diverse range of content.