PARIS runs end to end literature reviews with 15 agents
The point
PARIS, the PRISMA Agentic Research Intelligence System, tackles literature reviews, which stay slow because most tools address only single stages. A conductor agent in Claude Code orchestrates 15 specialized LLM agents, from scoping a question to exporting a manuscript draft.
Why it matters
For builders, persisted artifacts and human gates keep long multi agent workflows on track.
Key details
- Agents coordinate through artifacts the conductor validates before every transition
- QualityScore grades a corpus 0 to 100 before synthesis; one case rose from 74.6 to 81.9
- Two cases: 388 records, 100 included; 131 records, 123 included
Original ~3 min → 25 sec here Source: arXiv cs.AI, Nov 9, 2026