Webinar recording from May 6th, 2026: Build with the Amass API: live demos in Lovable, Claude Code, and Python — and a Q&A
Jörn and Alex from Amass walk through the newly opened Amass API for life sciences — what it is, why retrieval is the bottleneck for life-science AI agents, and how the hybrid BM25 + vector + re-ranker pipeline behind BioMedCore and TrialCore actually works.
Three live demos:
1) Vibe-coding an AI research assistant in Lovable in a single prompt
Plugging the Amass API into Claude Code as an installable skill
2) A ~150-line Python research agent built on the open amass-platform-starter repo
3) Plus a live Q&A on Python UI recommendations, pricing, how we curate clinicaltrials.gov data, and how we compare to providers like Sideline Intelligence.
Get started: https://platform.amass.tech — with starter credits
Chapters:
00:00 Intros (Jörn & Alex)
03:35 Why Amass: needle in the haystack + benchmark
08:43 Demo 1 — Vibe-coding an agent in Lovable
15:18 Under the hood: BioMedCore, TrialCore, retrieval engine
24:43 Demo 2 — Amass skill in Claude Code
31:31 Demo 3 — Python agent on the Amass API
46:17 Q&A
#LifeScience #AIagents #BiomedicalAI #RetrievalAugmentedGeneration #Amass #API