Mehmet Bartu

GPT Computer Assistant — Create intelligence for your products

Featured
121
•
Create your own intelligence specific to your product. With an agentic infrastructure, custom tools, and extensive API support, customize your intelligence in detail and establish extensive artificial intelligence infrastructures like apple intelligence.

Add a comment

Replies
Best
Mehmet Bartu
Maker
Launching soon!
đź“Ś
Hi folks, I'm Bartu, One of the co-founders of the GPT Computer Assistant (GCA). At OpenAI's GPT-4 launch in May, they said the future of their assistants was on new computers, and we wanted to democratize it. We designed the GPT computer assistant framework to support Windows, Linux, and MacOS.🧑‍💻 Why we say Intelligence, not Assistant - The Tool capability of the intelligence you will create with GCA is very high - Agentic infrastructure is available - You can add custom tools and create ordered tasks - We call it 'intelligence' because it can also be used as a framework. Assistants are usually tools that use the assistant capabilities of GPT-4 with an interface. When you allocate enough development time (700+ hours), you can reach GCA's level but we already offer it for free in a form that includes many APIs that you can easily customize. 🧩 We support both cloud and many local models. We believe that open-source models are critical for the democratization of AI. 👉 How to build intelligence for your product with the GPT Computer Assistant - A wide range of APIs are available for many things such as logo, product name, characterization - You can customize the agent infrastructure according to your product, so you can get higher quality AI outputs - By adding custom tools for agent infrastructure, you will get accurate results in specific tasks Experience✨ - It answers your question in less than a second in local models. - To get out of the monotony, we voice the answers the assistant gives in three different voices. - It understands when your conversation is over and acts according to your commands. - You can silence GCA without finishing its speech - You can see the path it followed to respond to your commands.
Shadman Nazim
@bartu Congrats! GCA is paving the way for truly intelligent assistants. A possible enhancement could be adding a learning mode where the AI adapts to specific user preferences over time. Wishing you great success with GCA!
Mehmet Bartu
Maker
Launching soon!
@dhruv_n1 Yes, that would be great. Thank you for the suggestion.
Onur ULUSOY
@bartu @dhruv_n1 Thanks, sure the new perspective of products will be this kind of interfaces.
Mehmet Bartu
Maker
Launching soon!
@dash4u We use it in the Upsonic codebase, we turned it into a copilot for ourselves, and it was especially useful in development.
VĂ­tor Soares
Congrats on the launch! where do you see CGA having the most common use cases? Explain it to me like if I wasn't a Dev ('cause I'm not ahah)
Mumtaz Vural
@vitor_pvsoares Ahahah thanks Vitor, A common use case could be for b2b SaaS companies to create intelligence for their software. Like a on-bording guys. :)
Mehmet Bartu
Maker
Launching soon!
@vitor_pvsoares For example, for a marketing tool, rather than just generating a message, you can ask it to research from forms and the web, thanks to the agent structure, this way you can get a more current result. You can ask it to write a blog about the topic it researches and have a much more successful text. After you build your own intelligence, you can add tool mechanisms that will enable the people using your product to onboard faster For e-commerce It can prepare a special report on what they can do to achieve more successful results by examining the Google analytics data of the people who open stores on the website. It can make price research, investigate the prices of the sold product in other stores, and give advice on why the shopping cart conversions are low. In our codebase product named Upsonic, we use many agents in topics such as writing code, checking code quality, setting code standards, passing the code through security, and auto-documentation of the code.
Onur ULUSOY
@vitor_pvsoares Thanks and for example you can create your ow Copilot for your b2b dev tools.
Joseph Abraham
@vitor_pvsoares Congrats on bringing your idea to life! Wishing you a smooth and prosperous journey. How can we best support you on this journey?
Mehmet Bartu
Maker
Launching soon!
@kjosephabraham We would be very happy if you could try the product and give feedback.
Alexandre Contador
This is so smart! Great addition to any AI infrastructure. Congrats on the launch
Mumtaz Vural
@alexdigitalmkt Hey! Nice to hear that, you can easily create a customize AI infrastructure for your product
Mehmet Bartu
Maker
Launching soon!
@alexdigitalmkt Absolutely, and it makes us happy to offer this to the community as open source.
Onur ULUSOY
@alexdigitalmkt Thanks and lets bring it to the future with Open Source Software community.
giorgi khatiashvili
Congrats,What level of technical expertise is required to effectively utilize GCA?
Onur ULUSOY
@giorgi_khatiashvili2 Thanks, and actually, if you can write a little bit of Python, you can create and customize your own intelligence for your products.
Mehmet Bartu
Maker
Launching soon!
@giorgi_khatiashvili2 We will share a tutorial soon.
Trisha Sinha
Congratulations on the launch! As a marketer, the use case of SEO analysis caught my eye. Can you share more use cases in terms of marketing?
Mumtaz Vural
@trisha_sinha20 Thanks Trisha, We have a lot of use cases about marketing and it relative with your imagination :) Especially in SEO you can get a specific recommendation from your Google Console
Mehmet Bartu
Maker
Launching soon!
@trisha_sinha20 You can save your data in Google Search Console to its memory, so it can give recommendations on where you need to focus to increase user acquisition and raise awareness about your product. In fact, for places where you think AI can be used in your product, we provide a structure that allows you to get higher quality results and enter ordered tasks and totaly free
Onur ULUSOY
@trisha_sinha20 Also the developers of the apps that you used for SEO can generate their own Intelligence to serve you. It can help in many areas with llm power.
Olivia Jane Mitchell
This looks intriguing, Bartu! I'm curious about how the GCA framework handles integration with existing systems. Is it straightforward to implement for developers who may not be deeply entrenched in AI or machine learning? Also, you mentioned that building with GCA can lead to high-quality outputs—what kind of metrics are you using to measure that? Any real-world examples or case studies of businesses benefiting from this? I’m especially interested in the flexibility of the agentic infrastructure you’ve offered. What kinds of custom tools have other makers found particularly useful? Looking forward to learning more!
Mehmet Bartu
Maker
Launching soon!
@oliviajanemitchell A little knowledge of Python is enough, you don't need to be an advanced developer. We say this thanks to GCA's agentic infrastructure and its ability to successfully complete ordered tasks. We currently have 23k downloads, but we haven't tracked which projects it's used in.
Kyrylo Silin
Hey Mehmet, How long does it typically take for someone to set up and start using GCA effectively? Could you give an example of a unique custom tool that can be added to the agent infrastructure? Congrats on the launch!
Onur ULUSOY
@kyrylosilin Thanks for your attention. Basic customization takes only 5 minute. Detailed one is just 30 minutes if your apps have API. And this is an example special tool: https://github.com/Upsonic/gpt-c...
Mehmet Bartu
Maker
Launching soon!
Hi @kyrylosilin You can set up GCA in a simple way. Think of adding a custom tool like configuring a specific task related to your business.
Fahrettin Cakir
Nice work and congratulations on the launch! Can we use internal user-level libraries of Windows using your product?
Onur ULUSOY
@fahrettin_cakir Yeah as you know the python is os-level language.
Elke
Wow, Bartu! This sounds like a game changer for custom AI solutions. The level of flexibility and control you’re offering with GCA is impressive. I mean, the ability to have an 'intelligence' framework instead of just a basic assistant really opens up new possibilities. I'm particularly intrigued by the voice responses—definitely adds a personal touch to interactions. Plus, supporting open-source models is a bold move for democratizing AI. Can't wait to see more makers jumping on board and building cool stuff with it! Upvote for sure! 🚀
Onur ULUSOY
@elke_qin Thanks and yeah the custom tools is so important. We also choose tool supported models for local to don't create any advantage for cloud.
Mehmet Bartu
Maker
Launching soon!
@elke_qin I love frameworks, they speed up the development process a lot. After we designed the GCA intelligence, we realized it could be a framework and we developed it for a long time. We hope it will become a popular framework, even if not as much as Django.
Ayhan Ergezen
Wow, that's a really clever approach. Congratulations.
Onur ULUSOY
@livelypencil Thanks to Open Source Community. <3
Mehmet Bartu
Maker
Launching soon!
@livelypencil Thank you
Max Strus
hey! i see you went with langchain, how do you like that so far? also groq as llm provider is cool, is it comparable with openai stuff?
Mehmet Bartu
Maker
Launching soon!
@promised_land Initially, we were having issues with local models, but they solved this problem later. Other than that, it was pretty good.
Onur ULUSOY
@promised_land We use; - langchain - openinterpreter - crewai and we support everyting with local support. Include TTS, STT and LLM models. GCA default one is really comparable with OpenAI desktop app but GCA is more than one shit app.
Tony Han
A few things that caught my eyes: it's open source, how raw the product interface is (this means you really focused on building this as a framework), the APIs are quite comprehensive, the local LLM support, and the omni experience. Often times though, the use cases are fairly nuanced, so a lot of customization has to be done to make it work. How much time would it take to fine tune the app? Congrats on the launch @bartu and team!
Mehmet Bartu
Maker
Launching soon!
@tonyhanded It's a very clever approach, Tony. A simple customization takes 5 minutes, but a detailed customization can take between 1 and 10 hours.
Antoine Gauthier
This looks like a super interesting project, @bartu! The concept of building an "intelligence" rather than a typical assistant is intriguing. I'm curious about the specific customization options available within the agentic infrastructure. What are the limitations, if any, when adding custom tools? Also, could you provide more details on how the local models differ in performance compared to cloud-based ones?
Onur ULUSOY
@bartu @antoine_gauthier_ Thanks, and as you said, the "intelligence" concept is quite different for us. The custom tools are Python functions for now. You can send a request to somewhere, or you can install a specific Python library as well. Actually, we just added doc string limitation. The local models improve themselves day by day. We are looking for new models every day. For today, the winner model is llama3.1 but still no vision capability. However, we have support for vision-capable model supports as well.
Ishwar Jha
Congratulations!!! I m certainly going to try it
Mehmet Bartu
Maker
Launching soon!
@ishwarjha Thank you. If you encounter a problem, you can always reach us through our Discord channel.
Onur ULUSOY
@ishwarjha Thanks and would be great to listen feedbacks afterwards.
Ema Elisi
Fantastic work, @bartu! The potential for custom AI tailored to specific products is amazing. With such robust API support and tools, it seems like Makers can really enhance their applications. Excited to see how this evolves and integrates into existing systems. Upvoted!
Onur ULUSOY
@bartu @ema_elisi Thanks, we believe to create a new kind of interface mindset.
Mehmet Bartu
Maker
Launching soon!
@ema_elisi Definitely, when makers want to design for their own products from scratch, they will spend a lot of time. We allocated more than 700+ hours to develop this. Instead, a maker can focus on more important matters.
dashuai
Great insights, @mumtazvural! The idea of custom intelligence for B2B SaaS is definitely a game changer. It can really enhance user onboarding and improve UX significantly. Excited to see how this evolves in the future! 🌟
Onur ULUSOY
@mumtazvural @ezshine Sure, also believing to Python community on our product also so nice.
Mehmet Bartu
Maker
Launching soon!
@ezshine The product will go to a very exciting place with the contributions of the open source community.
Mikhail Sidakov
That's nice, congrats on the launch!
Onur ULUSOY
@mikhail_sidakov Happy to be here.
Mehmet Bartu
Maker
Launching soon!
Ghost Kitty
Comment Deleted
Mehmet Bartu
Maker
Launching soon!
@zulkarnaim Thank you for kind wishes
Onur ULUSOY
@bartu @zulkarnaim Thanks and really nice to hear these words. We really like to create frameworks and infrastructures.
Congratulations on your launch! 🚀 This is a great solution 🙌
Mumtaz Vural
@mehmettalhasisman Thanks Talha, We're so excited to solve hard thingss
Onur ULUSOY
@mehmettalhasisman Thanks. thats really cool to hear.
René N
Uh i like the infrastructure part of this. Really exciting 🤩 Congratz on the launch 🚀🚀
Onur ULUSOY
@rene_nielsen_dk Thanks. We just well design our "Intelligence" concept.
Mehmet Bartu
Maker
Launching soon!
@rene_nielsen_dk Thank you🚀