NLP for edge inference | Offline nlp on a chip

Edge inference NLP for low spec devices

Infer text intelligently - On Device and offline !

Get in touch via email for a trial or other needs

Fast (<1 ms)

ThatNeedle can perform entity extraction, inference in less than 1 millisecond.

That's almost 10x faster than most other options in the market and makes ThatNeedle suited for real-time NLP

Domain intelligence

Efficient use of domain specific cognition-graphs for domain specific applications results in deeper and better inference.

Hardware accelerator not required !

Will work on commodity CPUs and SoC also. We are not dependent on expensive hardware accelerators and GPUs

Portable & Compact

Compatible with most low spec embedded systems and chip architectures.

Testimonial

testimonial text

Founder & CEO

Tech Market Insight

ThatNeedle team truly understood our requirements and came up with a very high quality output in a very quick time. The implementation was easy with absolutely no integration issues. We are getting higher click through rates and more page views after implementing the new search. We are now working on next phase to make the search truly semantic in every sense.

Vivek Agarwal

Founder: realizing.in

Vivek Agarwal

Founder: realizing.in

Setting up the microservice on GNU/Linux flavours

1. After you have downloaded the binary you can run it as :-
thatneedle_topic_for_tech.l <port_number>

2. You will be prompted for a license key
Please type it and press enter. Enter license key here: your-licence-key
Remember that you must be connected to the internet for this.

3. Once the is accepted and validated, you are ready to use the service on your server !
Ready.. You can access the service at http://:6060/?q=your technology text here

Calling the microservice from a python program

Topic extraction in python made easy with our microservice

import requests import json import urllib query = "Setting up Tensorflow on linux" # from the user/ UI query = urllib.parse.quote(query) results = requests.get(F"http://localhost:9090/?q={query}") results_dictionary = json.loads(results.text)

About the deliverables:

1. Deliverable is a binary for 1 machine ( single server license )

2. Please note that source-code is not delivered with this transaction.

3. You can run this as a microservice on your server.

4. You can use it in accordance with EULA(End user license agreement)

5. Please ask for an online trial before you purchase. There are no refunds.