I run my software agency named Overflow Labs. Besides offering software consultancy, I ran a few other in-house software products. A big part of my time goes into the SEO, branding and marketing of Overflow Labs and these portfolio websites.
As time passed, I saw a clear need to automate my SEO and marketing workflows to avoid spreading too thin. That’s when I decided to put together what I’ve been building in-house and share it with the world — after all, if this project gets more funding from people using it, I could funnel more resources into its development and we all benefit. That’s how Fewlogs started.
Fewlogs is the public deployment of the platform we use at Overflow Labs to enhance any AI prompt-based workflow. It can be used with any language model or writing tool that uses prompts, so we make sure to adapt to any future AI developments.
We want Fewlogs to store all the expertly curated sources regardless of the format, compile a series of writing styles that we can replicate and keep our best-performing prompts in private repositories.
Content generation is important, but not the only priority. We want Fewlogs to be open to its internals so that we can leverage its semantic capabilities to perform content research and discover new insights from large content databases.
Linking and content research is so much easier with a semantic search tool that can work with huge amounts of content. This has been a huge win.
Content generation is another obvious point. We all love ChatGPT but its output is too generic, too ChatGPT. We built Fewlogs to leverage Retrieval Augmented Generation (RAG) which allows for embedding informational and styling context so that the generated output is tailored exactly to what we want.