Infodrops needed a research engine and the application around it. Ego Eimi built the research loop, authenticated jobs, progress updates, report exports and subscriptions as a connected SaaS product.
Deep research is slow because one search never settles it. The work is deciding what is relevant, extracting what matters, finding gaps, and following those gaps until the answer is complete.
Infodrops needed to behave like a patient analyst: explore, score, extract, follow up, and return a useful brief instead of a pile of links.
The application used Next.js and React for the interface, with Fastify handling authentication, jobs, orchestration and Stripe subscriptions. Users could follow a research job’s progress and export the resulting report.
The research pipeline connected iterative search, relevance scoring, extraction and follow-up exploration. Firecrawl supported search and crawling, a Python service handled keyword clustering and sentiment, and Supabase Postgres stored application data.
The delivery included both the research engine and the application needed to operate it: job state, administration, billing, caching, exports and deployment. Research became an inspectable workflow in a product rather than a background script with only a final response.
For an AI SaaS, the business depends on job completion, provider availability and billing as well as model output. The operating record needs to explain retries, cached results, customer access and deployment. These are the dependencies a technology team inherits when it takes responsibility for the product.
Project contribution: Ego Eimi. The scope described above identifies the work behind this case.
Editorial review: Ego Eimi · Updated September 6, 2026 · About the team
Your technology
Start with the business
Tell us which system matters, what has changed and who controls it. We will define a reviewable next step. Read Foundation and the commitments and evidence.
