---
type: transcript
date: 2025-06-11
youtube_id: gQb9VCrBS2w
---
# Transcript: Introducing Fire Enrich: An Open-Source Clay Alternative with AI-Powered Data Enrichment
I'm super excited to announce Fireenrich. This is an open source project that I've been working on. And what it allows you to do is you can drag and drop a CSV. The first thing that it's going to do is it's going to autodetect the email columns. You want to make sure that you have a header as well as all of the emails that are all valid and all of that so you don't run into any issues. Once you have a list of different emails, what it's also going to do is it's going to filter out a number of different common providers. You can think of things like Google or Proton Mail, Yahoo, so on and so forth. It's going to remove that. And one thing that I want to know with the system is it really biases towards being good at finding information about companies. Right now, once we've determined what the email column is, we can select from a number of different fields that we want to enrich. Things like company name, description, industry, employee count, year founded, so on and so forth. Just to touch on how these different fields work on the back end. Each different pill maps to a different agent strategy. And right now there's a handful. For instance, if you're looking up a company's profile, there's going to be an agent that biases towards a strategy to find that particular information about the company. If it's looking for funding information, again, it's going to bias towards different sources. And the way that I wanted to set this up is in a way where it's easy to extend this. Say if you want to add in other agents or if there's particular things or proprietary APIs that you want to incorporate within this, you could break that out within an agent and have the orchestrator just go and find and leverage the different sources or strategies for each different agent. Let's just run through an example. What I'll do is I'll remove let's say company description and I'll add funding raised as well as funding stage. Then from there, there is also an agent that will go and if you have unique queries or specific things, it's going to just try and search the internet and extract that information based on whatever you put within here. Say if I put something like CEO name and I add that in. What this is going to do is we're going to generate the structure for the schema and we can go ahead and accept that here. And from here we can go ahead and start the enrichment process. The way that this is broken out, we have different sub agents that focus on particular tasks. For a company profile, it will go and bias towards a strategy that's focused particularly on trying to find information about say the company name or description. Whereas say if we ask for funding information that is going to be a separate agent that is broken out in a really clean way where even if you want to add in different prompting strategies or even different tools to have those different agents leverage different things like say maybe you want to add in a financial API or something like that. You you can definitely do all of that sort of stuff. If I just go through, you can see all of the detailed information for all of the different rows. Say if you have 10 different columns and it's a little bit hard to see all of the different values, you can go ahead and expand that. The other thing that I want to point out is you can go and you can export the CSV at any point. Say if you have a 100 different rows that it's going through at any point if you want to have a checkpoint, you can go ahead and download that. Now, additionally, we do also have a list of different skipped emails for common providers like Yahoo or Hotmail, it will group all of those together. And then you also have the option to download it as JSON. The other thing that is set up within here is when you click through to the sources, it will also show that corroborating piece or pieces of information of where it found that information for the CEO column. Within here, I can see we have Rocket Reach and that's the website that I leveraged to gain all of the information. It is live on GitHub now. I also added in a really extensive readme that you can go and check out. I also added in a mermaid diagram of the agent architecture so you can see exactly how all of the different phases work where you can see depending on the different queries all of the different strategies that are set up. Just quickly to dive into both the tech stack as well as how the agent orchestration is set up within here. What we have is we're leveraging the OpenAI agents SDK which just came out. And what's really nice with this SDK is we can initialize these clean agents and ultimately have them coordinate together within the orchestration process. Within here, we have the company profile agent. We have the prompting and instructions for us. And within each of the agents, we have all of the different tools that we can leverage. You can go ahead change these out, make your own custom agents, add in your own custom tools. But otherwise, that's pretty much it for this video. I definitely appreciate a star on the repo. If you found this video useful, please comment, share, and subscribe. Otherwise, until the next
