A command-line tool for NautilusDB, a fully-managed, cloud-native vector search service.
NautilusDB is currently in public alpha. We're actively improving the product and releasing new features and we'd love to hear your feedback! Please take a moment to fill out this feedback form to help us understand your use-case better.
By default, all collections are subject to permanent deletion after 2 weeks. Please let us know if you need to keep it for longer via the feedback form.
You can try out NautilusDB right away. We have prepared a special public
collection openai-web
that can answer questions about the contents of www.openai.com
nautiluscli ask openai-web "what is red team?"
Install a released NautilusDB cli from pip.
python3 version >= 3.10
is required
pip3 install nautiluscli
You can create a new API key and set it in NAUTILUSDB_API_KEY
environment variable. NautilusDB will use this key to authorize access to
collections.
nautiluscli create-api-key
export NAUTILUSDB_API_KEY='<key>'
nautiluscli info
Create a new collection myCollection
in the shared demo account.
A collection will be created that is only accessible to the configured API key.
nautiluscli create-collection myCollection
See the list of Collection accessible to you, including myCollection
that you just created.
nautiluscli list-collections
Index a PDF into myCollection
. In this example, we will index the original research paper on Transformers.
nautiluscli upload-file myCollection https://arxiv.org/pdf/1706.03762.pdf
Alternatively, upload a PDF from local file system for indexing.
nautiluscli upload-file myCollection README.md
You can now ask questions within the context of a collection and get answers.
nautiluscli ask myCollection "what is a transformer?"
You can optionally delete the collection. Deletions are non-reversible.
nautiluscli delete-collection myCollection