AIAny
AI Agent2026
Icon for item

PaperOrchestra

Turns research ideas and experiment logs into submission-ready LaTeX papers, coordinating literature review, figure generation, and revision. Its agents verify citations and refine drafts using simulated review feedback.

Introduction

Automating experiments is only part of AI-assisted research: turning scattered notes and results into a coherent paper is a separate, difficult bottleneck. PaperOrchestra focuses on that handoff, coordinating specialist agents around a researcher’s existing materials rather than requiring a fixed experimental pipeline.

What Sets It Apart
  • Literature discovery is paired with Semantic Scholar verification, so citations are checked against paper records instead of being accepted solely from generated text.
  • Outline, literature review, plotting, and section writing are coordinated as distinct stages; plotting and literature work can run in parallel. This makes the workflow easier to adapt to existing research inputs and conference templates.
  • A simulated-review loop accepts revisions only when evaluation scores improve or meet its tie-breaking criteria. In reported side-by-side human evaluations, it achieved a 50–68 percentage-point win-rate margin on literature-review quality and 14–38 points on overall manuscript quality against autonomous-writing baselines.
Who It Suits

Great fit for researchers who already have an idea, experiment logs, and results, and want help assembling them into a cited, formatted manuscript. Look elsewhere if you need an agent to conduct and validate the experiments themselves, or if you need a complete out-of-the-box dataset: the paper-generation dataset is not included in the repository, and the workflow depends on external model and literature-search services.

Information

  • Websitegithub.com
  • OrganizationsGoogle Cloud AI Research
  • AuthorsYiwen Song, Yale Song, Tomas Pfister, Jinsung Yoon
  • Published date2026/05/01

Categories

More Items

GitHub
AI Agent2026

Adds AI agents to a self-hosted email, calendar, and file workspace for searching mail, drafting replies, and managing events. It also connects to other AI tools over MCP, with infrastructure and mail configuration left to the operator.

GitHub
AI Agent2026

Build, share and govern company AI agents on infrastructure you control, connecting them to internal documents and tools. Roles, budgets, approvals and run records add operational oversight; hosted and local models are supported.

GitHub
AI Agent2025

Streams schema-validated UI descriptions from agents and renders them as native widgets across web and mobile frameworks. Its data-only protocol enables incremental updates without letting agents send executable UI code.