Ar9av/PaperOrchestra

An automated AI research-paper writer based off Google's PaperOrchestra paper's implementation through a skills - benchmark + autoraters using any coding agent (Claude Code, Cursor, Antigravity, Cline, Aider). No API keys, no LLM SDKs.

What it solves

PaperOrchestra automates the process of turning unstructured research materials—such as experimental logs, raw ideas, and scattered agent chat histories—into a submission-ready LaTeX research paper. It addresses the quality gap between single-agent writing and professional academic standards by implementing a structured, multi-agent pipeline.

How it works

Instead of being a standalone application, this project is a "skill pack" designed to be plugged into existing coding agents (like Claude Code, Cursor, or Aider). It provides the prompts, schemas, and deterministic helper scripts that guide a host agent through a five-step pipeline:

  1. Outline: Creates a structured JSON plan for plotting, literature review, and sections.
  2. Plotting: Renders plots and conceptual diagrams (optionally using PaperBanana).
  3. Literature Review: Searches for candidates, verifies them via Semantic Scholar, and drafts the related work.
  4. Section Writing: Drafts the remaining sections and integrates tables and figures.
  5. Content Refinement: Performs simulated peer review and refines the text based on strict halt rules.

An optional Research Aggregator can be run first to scan agent caches (e.g., .claude/, .cursor/) and synthesize scattered logs into the structured inputs required by the pipeline.

Who it’s for

AI researchers and developers who use coding agents to conduct experiments and want to automate the drafting of their academic papers while maintaining high fidelity to research data.

Highlights

  • Host-Agent Agnostic: Works with any agent capable of running local scripts and performing web searches; requires no internal API keys.
  • Paper-Fidelity: Prompts are reproduced verbatim from the original arXiv:2604.05018 research paper.
  • Deterministic Helpers: Includes scripts for BibTeX formatting, JSON schema validation, and LaTeX sanity checks to ensure structural correctness.
  • Integration Ecosystem: Supports optional enhancements via Semantic Scholar (for citations), PaperBanana (for high-quality diagrams), and Exa (for specialized research search).

Related

  • Project
    aiming-lab/AutoResearchClaw

    An autonomous, collaborative research pipeline that turns a research idea into a conference-ready academic paper with real citations and sandbox experiments.

  • Project
    johnson7788/MultiAgentPPT

    A multi-agent system that automates the generation of high-quality, editable PPTs by coordinating specialized agents for outlining, parallel research, and quality checking.

  • Project
    CRui5in/paper-ppt-agent

    An AI-powered tool that transforms academic papers (PDF/TeX) into editable PowerPoint presentations using a multi-agent pipeline for content extraction and visual design.

  • Project
    FeijiangHan/PaperForge

    A structured system prompt and skill framework that transforms LLMs into active paper-reading assistants to reconstruct author reasoning and generate research ideas.

  • Project
    claesbackman/AI-research-feedback

    A collection of Claude Code skills that provide multi-agent academic reviews for research papers, grant proposals, and pre-analysis plans, including paper-code reproducibility checks.