Digest Engine: Indie Author & Publishing News

Inspiration

We're built around independent publishing — writing, narrating, and distributing work across formats, which means also being the marketing department, the trend-watcher, and the industry analyst. Staying current on the indie publishing landscape used to mean a dozen open tabs: Jane Friedman's blog, Author Media, The Creative Penn, agent Q&As, retailer policy changes — all worth reading, none of it in one place. We built Digest Engine because we were tired of doing that manually, and because it was exactly the kind of messy, recurring chore an autonomous agent should own outright.

What it does

Digest Engine is a background agent, not a chatbot. It oversees a news collection module, pulling headlines from approved RSS sources relative to frequency. When a new story is detected, Gemini 3.6 uses the URL Context feature to examine the snapshot of recent news articles to generate a new digest.

How we built it

The agent runs on Gemini 3.6 Flash, built to operate outside a standard chat loop — it executes on a schedule, handles the retrieval and curation pipeline itself, and writes directly to the page. Our stack is gathering data via RSS and other methods, hosting, and handling scheduling mechanisms. The front end is a straightforward page on an existing author site, michael-finney.com, styled to match the rest of the platform rather than living as a separate product.

Jfin was responsible for backend/agent building and testing - all heavy technical load

Michael was responsible for content/media, frontend, design, FTP/hosting management

Challenges we ran into

The largest building challenge we faced was integrating these new Gemini-powered curation features into the existing website's architecture in a way that preserved 100% uptime. This was accomplished with cron job-based atomic-state caching, meaning this complex system could mesh easily into the existing hosting infrastructure and actively deployed website. We kept all the heavy processing on Google Cloud, separate from the web host.

Accomplishments that we're proud of

A working agent that runs unattended and has already been live-updating through the build — the page updates itself twice daily and the entries are real, dated, and sourced. We have also been getting great feedback from the indie author and publishing community, who we have been sharing the progress with over the last week. Not only are people visiting but they are returning and clicking through to read the articles the page has sourced.

What we learned

We learned that with focus and effort, AI agents can enable a brand new type of web page that is engaged with today's web and active in being thoughtful about industry topics that can be incredibly difficult to keep up with otherwise.

What's next for Digest Engine: Indie Author & Publishing News

Our next goal is to increase the number of sources we are pulling in. We also intend to create a more robust and dynamic backend that will enable users to deploy multiple feeds (various industries/topics, or eveninclusions/exlusions for a single source to be split into multiple digest feeds on multiple pages). Additionally, we intend to test this by creating a page feed of opportunities for writers and authors and then a news feed for indie filmmakers. Further, we are considering how this tool could perhaps become a white-label solution that others can deploy and customize)