Ollie Martynov Profile picture
Apr 4, 2023 โ€ข 10 tweets โ€ข 4 min read โ€ข Read on X
๐Ÿงต1/8 Here is how to build a pipeline that automates the process of creating, researching, improving & finalizing an article on any topic. I used #langchain & @OpenAI, Google Search and Google News API! Let's dive in!๐Ÿ‘‡ Image
๐Ÿงต2/8 The pipeline starts by loading news articles on a chosen topic. In our example it is "Canada budget". It then creates a summary of these articles, which will be used as the basis for our first draft. Image
๐Ÿงต3/8 The first draft is generated by using custom author and question chains in a sequential manner. The author chain creates a coherent narrative, while the question chain generates relevant questions to research further. #langchain #OpenAI
๐Ÿงต4/8 After generating the first draft, it's time to find answers to the questions raised. The pipeline leverages Google searches to obtain relevant information for these questions. Image
๐Ÿงต5/8 With the answers in hand, the extra information is added to the initial draft. This forms a more comprehensive article with a wider range of insights. #ContentCreation
๐Ÿงต6/8 To polish the content further, the pipeline uses custom writer and editor chains sequentially. The writer chain adds depth to the article, while the editor chain refines the content, structure and format. #AIWriting
๐Ÿงต7/8 Finally, the pipeline outputs both a final draft and a completed article. The final draft contains all the revisions and improvements, while the final article is ready to be published. #Automation
๐Ÿงต8/8 By combining #langchain, @OpenAI, and Google Search API, this pipeline streamlines the process of generating high-quality articles on any topic. For example we are able to transform "Canada budget" query into the following .md file: Image
p.s. Liked this thread?
Please subscribe for more awesome content like this. ๐Ÿ’ชโœ…
Many have reached out to me about the full source code for the pipeline. Its quite messy but its works.
๐Ÿฅฒ
github.com/olliethedev/auโ€ฆ

โ€ข โ€ข โ€ข

Missing some Tweet in this thread? You can try to force a refresh
ใ€€

Keep Current with Ollie Martynov

Ollie Martynov Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!

PDF

Twitter may remove this content at anytime! Save it as PDF for later use!

Try unrolling a thread yourself!

how to unroll video
  1. Follow @ThreadReaderApp to mention us!

  2. From a Twitter thread mention us with a keyword "unroll"
@threadreaderapp unroll

Practice here first or read more on our help page!

More from @olliethedev

Jul 7, 2023
Did you know that the fusion of AI agents and Bitcoin has the potential to revolutionize our world? ๐ŸŒ Let's take a deeper dive into this fascinating confluence of technology. ๐Ÿ‘€๐Ÿ‘‡๐Ÿฝ Image
AI agents are programmable entities interacting with their environments to achieve set goals. They can adapt to new situations and perform a variety of tasks, from debugging code to making API calls for purchasing data or real-world resources. ๐Ÿค–๐Ÿ’ป
But how do these AI agents pay for resources? With an innovative protocol called L402, AI agents can now leverage Bitcoin - internet's native currency - to pay for web content access directly. A digital wallet for your digital assistant, if you will.๐Ÿ’ฐ๐ŸŒ
Read 12 tweets
May 31, 2023
๐Ÿค– Ever wished you could teach an AI to mimic website designs you love? Here's how we can make it possible with the @LangChainAI library and a dash of GPT magic.โœจ Learn how to build your own AI agent that loves to plagiarize. ๐Ÿ‘‡
๐Ÿ’ป The core of our project is a simple NodeJS and TypeScript app. We create a prompt specifying the site we want our AI to learn styles from. And provide the Style Tool available to use for this task. Image
๐Ÿ”ฌ How does the Style Tool work? Our secret sauce is the JSDOM library. It helps us render the page headlessly and parse through elements like a pro. Here is a video explaining the details. The final webpage is inspired by the styles of @itsPaulAi's answera.ai.
Read 8 tweets
May 17, 2023
๐Ÿš€ The customer service industry, valued over $10B, is ripe for AI-driven transformation. With @LangChainAI and @awscloud, automate your customer support with a few lines of code. Let's dive in! #AI #CustomerService
Our AI-agent acts as an automated customer support rep, intercepting emails sent by users. It understands the user's issue, gathers necessary information, and crafts a comprehensive response. The result? A seamless interaction, quick resolution, and improved customer experience. Image
@LangChainAI is a powerful library for AI-powered applications. Our customer support system revolves around its robust features, which simplify the development process. Check out how it powers our AI-agent. Image
Read 8 tweets

Did Thread Reader help you today?

Support us! We are indie developers!


This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3/month or $30/year) and get exclusive features!

Become Premium

Don't want to be a Premium member but still want to support us?

Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal

Or Donate anonymously using crypto!

Ethereum

0xfe58350B80634f60Fa6Dc149a72b4DFbc17D341E copy

Bitcoin

3ATGMxNzCUFzxpMCHL5sWSt4DVtS8UqXpi copy

Thank you for your support!

Follow Us!

:(