Scribe Earns a 39 Proof of Usefulness Score by Building an Agent that Synthesizes Group Chat Links into High-Signal Digests

In this interview, we sit down with Justin Irabor to discuss Scribe, an intelligent agent that listens to group chats, captures shared links, and synthesizes them into highly relevant, scheduled digests. Currently in its alpha stage on Telegram, Scribe aims to solve the widespread problem of information overload and missed resources in active group communications.

What does Scribe do? And why is now the time for it to exist?

Scribe is an agent that listens in group chats for links, and synthesizes them into a high-signal digest sent on a schedule. It was inspired by me and my friends always sharing links to interesting things, with us missing most of the links because it's tough to keep track of that information. Scribe is currently available on Telegram (in alpha). Now’s a good time for Scribe to exist because our primary mode of knowledge sharing has shifted to fast-paced group chats, making it critical to have tools that reliably capture and preserve valuable insights before they get lost in the feed.

How many people does Scribe reach?

10 per month (we are currently in alpha)

Who does your Scribe serve? What’s exciting about your users and customers?

Small, close-knit groups that share interesting information that can be curated into high-signal, low-noise digests people can subscribe to to keep updated. Think discord channels, etc.

What technologies were used in the making of Scribe? And why did you choose ones most essential to your tech stack?

We built Scribe utilizing Neo4j, Bright Data, and Algolia as our core technological foundation. Bright Data provides a robust web scraping infrastructure required to reliably extract metadata from varied shared links, while Neo4j allows us to map the complex, graphical relationships between chat members and the information they share. Finally, Algolia ensures that all gathered intelligence can be rapidly queried and searched when synthesizing our low-noise digests.

What is the traction to date for Scribe?

Since Scribe is still currently in its alpha stage, there is no public evidence of traction just yet. Our focus has been entirely on iterating the core web scraping features and relationship mappings with a closed circle of early internal testers before scaling up our public presence.

What excites you about Scribe's potential usefulness?

Two things:

1. Our internal users are enthusiastic about the quality of the digests they generate, and note that it'll allow them to get more value from group chats.

2. The technical challenges of a robust, sturdy web scraper and relationships mapping have been...revealing.