CEO of Inspire Search, Tsubasa Kato was invited to NVIDIA AI Summit Japan 2024 held in Tokyo, Japan.

CEO of Inspire Search, Tsubasa Kato was able to see Jensen Huang on stage on the 13th, and was very happy about it. He could feel the energy flowing through the AI, Gen AI, LLM, GPU, GPGPU sector in through this event. He is very thankful to NVIDIA for the opportunity.

Update on the LLM based web crawler (NVIDIA NIM Version)

Another alpha version of LLM Crawler is now made. We used NVIDIA NIM and Llama 3.1 405b as model.

Update on the LLM based web crawler

The alpha version of LLM Crawler is now made. We used Groq and Llama 3 8b 8192 as model. You can try the demo from: http://www.ideate.vision:5001

Update on the LLM based web crawler

We have created an improved version of a LLM based web crawler using Llama 3 Instruct. The demo is running on NVIDIA RTX 3090.

Click here for a demo running on Apple M1 Pro.

We are developing the LLM based web crawler further.

We have created a very early alpha version of a LLM based web crawler using Llama 3 Instruct. This demo is running on NVIDIA H100.

We upgraded the GPT Assistant for Grow My Search.

GPT Assistant for Grow My Search has been updated to GPT-4o, the latest model from OpenAI. Here is a video of it in action.

We experimented with Meta's Llama 3 Instruct that was released on April 18th, 2024.

Here is a video of us experimenting with the Llama 3 on NVIDIA RTX 3090.

New:Here is a video of us experimenting Llama 3 on Ollama using NVIDIA TESLA T4

We are experimenting with LLM and XLM such as Google's gemma-2b, gemma-7b and xlm-roberta-ner-japanese based on FacebookAI's XLM-RoBERTa.

We are experimenting with rather lightweight models to see if we can use it to guide web crawlers and find different ways to traverse the web graph.
Most recent: gemma-2b + Beautiful Soup web crawling sample on H100 (80GB VRAM, 200GB RAM)
Test video of gemma-7b on GeForce RTX 3090 (24GB VRAM, 64GB RAM)
Test video of xlm-roberta-ner-japanese on GeForce RTX 3090 (24GB VRAM, 64GB RAM)

We are experimenting with new LLMs with M1 Pro and GeForce RTX 3090 and inference API.

Inspire Search Corporation is now experimenting many LLM (Large Language Model) based test projects to enhance search and NLP projects. We recently experimented with Microsoft's Phi-2, Mistral's AI API, and more projects using Ollama. Keep tuned for more updates!
Test video of Microsoft's Phi-2 on GeForce RTX 3090 (24GB VRAM, 64GB RAM)
Test video of Microsoft's Phi-2 on M1 Pro (16GB memory)

Inspire Search Corporation is now a part of Microsoft for Startups Founders Hub

Inspire Search Corporation is now a part of Microsoft Startups Founders Hub. We will be utilizing the Azure Cloud and more to improve our products.

CEO attended LLM course in AIST

Tsubasa Kato, the CEO of Inspire Search Corporation, attended LLM (Large Language Model) training course in AIST (National Institute of Advanced Industrial Science and Technology) in Tokyo, Japan on 10/23/2023. At the training course, the participants were able to use ABCI's NVIDIA V100 cluster to fine tune the LLM.

Demo of GPT-4 Turbo AI Assistant for Grow My Search

We made a demo video of GPT-4 Turbo being used as AI Assistant for Grow My Search.

Concept for Container-based Mobile Search Appliance.

As there are wars and disasters in the world, your infrastructure may not always be secure. It may be damaged to a critical level. Inspire Search proposes the concept of a solution where the search appliance can be moved on a container basis and still gather information from a safe location.

Discussion with Google Bard on how to expand the LLM based seed expander
Tsubasa Kato had a discussion with Google's Bard on how to expand the LLM based seed expander - web crawler source code.
Take a look at it here:
bard-llm-webcrawler-discussion.pdf

LLM based seed expander - web crawler source code is on GitHub!

We are developing a LLM based seed expander - web crawler. We have released the source code in development.

Language: Python 3 https://github.com/stingraze/llm-seed-url-generator

Tsubasa Kato (CEO of Inspire Search) wrote an article on medium explaining what the LLM based seed expander - web crawler does.
Visit here for the article: https://stingraze.medium.com/llm-based-web-crawler-9e79fdbf7d0b

Grow My Search for Android is out on Google Play!

We have released Grow My Search for Android on Google Play! It is available for free.

New Advertisement for Grow My Search App is Out!

New brochure Made!
We uploaded a new brochure for Grow My Search / Inspire Search. (PDF) Download here.

We updated the Portfolio Page.

We added the demo of a web crawler that uses NVIDIA GPU with BERT, PyTorch and CUDA.

Go Check it Out!

We have been accepted to the Yext Startup Program!

Inspire Search is delighted to announce that we have been accepted to Yext Startup Program. We will be providing a new experience for our search products. Please stay tuned!

Metaphor.systems is added as a source search engine on Grow My Search

The Grow My Search app for iPhone / iPad / Apple Silicon is now free to download! We will really appreciate if you can send feedback!

We released Grow My Search App on Apple's App Store!

Tsubasa Kato, Inspire Search Corporation's CEO was interviewed by ANI News. Watch the video on YouTube here.

Tsubasa Kato, Inspire Search Corporation's CEO attended G20 - B20 in Imphal, India as a delegate from Japan from February 17th to 19th.

Tsubasa Kato, Inspire Search Corporation's CEO has been awarded an US patent. (US 11,556,602 B2)
Go to USPTO site

We are currently developing Grow My Search for Mac. It will be released to the App Store soon.
Here is the video of the app in development. (YouTube)

We launched Grow My Search on Product Hunt! Visit us here: