# PaperOrchestra: Google Research's Multi-Agent LaTeX Paper Writing Framework

> Source: <https://explainx.ai/blog/google-paper-orchestra-multi-agent-research-writing-github-2026?utm_source=rss&utm_medium=feed&utm_campaign=blog>
> Published: 2026-11-16 00:00:00+00:00

**PaperOrchestra is Google Cloud AI Research's open-source multi-agent pipeline that turns experiment logs and bullet-point ideas into conference-formatted LaTeX — with automated literature review, section writing, refinement, and optional plotting.** The [google-research/paper-orchestra](https://github.com/google-research/paper-orchestra?ref=explainx) repo listed **1,034 GitHub stars** when we fetched it on **November 16, 2026**, alongside an **[arXiv:2604.05018](https://arxiv.org/abs/2604.05018?ref=explainx)** citation in the README.

For ML engineers, it is a **concrete reference architecture** for decomposing "write the paper" into **orchestrated agents** — not a substitute for peer review, human authorship ethics, or **[replication culture](https://explainx.ai/blog/rope-positional-embeddings-modern-llms-raschka-guide-2026)**.

## TL;DR

| Component | Role | 
|---|---|
| **Outline agent** | Structure from raw notes | 
| **Literature agent** | Search, prioritize, integrate citations | 
| **Section writers** | Draft body text per section | 
| **Refinement agent** | Iterative polish | 
| **Plotting agent** | Optional code-generated figures | 
| **Output** | LaTeX under `templates/` (e.g., CVPR 2025) | 

## Why Google open-sourced another agent framework

Research teams already use **[Claude for lattice work](https://explainx.ai/blog/edward-witten-claude-opus-5-5-lattice-chiral-fermions-arxiv-2026)** and **[GPT-6 Astra science blogging](https://explainx.ai/blog/openai-gpt-6-astra-standard-model-lattice-nathaniel-craig-science-blog-2026)**. The bottleneck shifted to **manuscript assembly**: related work, consistent notation, figure regeneration after ablations.

PaperOrchestra attacks that **last mile** with **specialists** instead of one monolithic prompt — the same pattern as **[Sakana multi-agent self-supervision](https://explainx.ai/blog/sakana-mass-multi-agent-self-supervision-recursive-self-improvement-2026)** but aimed at **LaTeX delivery**.

The README credits **Yiwen Song, Yale Song, Tomas Pfister, and Jinsung Yoon** at **Google Cloud AI Research**. The project site is [yiwen-song.github.io/paper_orchestra/](https://yiwen-song.github.io/paper_orchestra/?ref=explainx).

## Pipeline architecture

A large cream ring routes three small green parcels outward — PaperOrchestra's specialized writer, literature, and refinement agents

PaperOrchestra advertises five coordinated stages:

1. **Outline generation** from unconstrained pre-writing material.
2. **Literature review** with search, prioritization, and integration into prose.
3. **Section writing** aligned to the outline.
4. **Content refinement** passes.
5. **Plotting** (optional) from executable code.

A **Streamlit demo** lives under `frontend/` for interactive runs — useful for **lab demos**, not production CI.

Compare to **[TradingAgents' multi-role finance simulators](https://explainx.ai/blog/tradingagents-multi-agent-llm-trading-framework-v0-7-chatgpt-plan-hn-skeptics-guide-2026)** — both show **role boundaries** beat one chat thread, but PaperOrchestra targets **static PDF artifacts**.

## Installation and environment

```
conda create -n paper_orchestra python=3.11
conda activate paper_orchestra
pip install -r requirements.txt
```

Create `.env` with keys the README lists:

- `SEMANTIC_SCHOLAR_API_KEY`
- `OPENAI_API_KEY` and/or`GEMINI_API_KEY` or Vertex (`VERTEX_AI_PROJECT` ,`VERTEX_AI_LOCATION` )
- Optional SMTP fields if the demo emails finished PDFs to you

**Dataset gap:** README explicitly says **evaluation data is not bundled** — future release under `datasets/`. You supply **`--raw_materials_dir`** with your own logs.

## CLI usage

Bash wrapper:

```
./paper_writing_cli.sh
```

Direct Python:

```
python paper_writing_cli.py \
  --raw_materials_dir /path/to/raw/ \
  --latex_template_dir /path/to/templates/cvpr2025/ \
  --output_dir /path/to/output/ \
  --use_plotting true
```

Key flags:

- **`--writer_model_name`** — LLM for drafting and literature agents.
- **`--reflection_model_name`** — LLM for critique passes.
- **`--use_plotting`** — enable figure agent (default false).

Templates live under `templates/`; add a conference by creating a new subfolder with the `.sty` and skeleton TeX your venue requires.

## Repository layout for contributors

| Path | Purpose | 
|---|---|
| `methods/` | Agent logic and orchestration | 
| `autoraters/` | Automated evaluation scripts | 
| `utils/` | PDF helpers, API utilities | 
| `templates/` | CVPR and other conference shells | 
| `frontend/` | Streamlit UI | 

`autoraters/` signals Google intends **measurable quality gates** — align your internal adoption with **[Nvidia SkillEvaluator thinking](https://explainx.ai/blog/nvidia-skillevaluator-agent-skills-three-tier-cli-guide-2026)** for agent outputs.

## Ethics, authorship, and review risk

Automated manuscripts raise **venue policy** and **integrity** questions:

- **Disclose LLM assistance** per journal and conference rules.
- **Verify citations** — retrieval agents can**hallucinate papers** or misquote abstracts.
- **Do not submit unreviewed LaTeX** — treat output as**draft zero** , not camera-ready.

Our **[AI-generated code license debate](https://explainx.ai/blog/ai-generated-code-open-source-license-debate-2026)** applies doubly when generated text ships in **Apache-licensed repos** with human names on the byline.

PaperOrchestra is **not** [OpenAI's Orchestra Bio drug agents](https://explainx.ai/blog/openai-acquires-orchestra-bio-rosalind-drug-development-agents-2026) — naming collision only.

## How it fits the multi-agent research stack

| Layer | Example in ecosystem | 
|---|---|
| **Experiment** | Hy4, Qwen, internal baselines ( [frontier comparison](https://explainx.ai/blog/frontier-ai-models-compared-fable-5-2-gpt-6-astra-gemini-4-carbon-grok-4-7-kimi-k3-1-deepseek-2026) ) | 
| **Benchmark** | [VibePrevisBench gaming agents](https://explainx.ai/blog/vibeprevisbench-kling-blender-agent-4d-scene-benchmark-2026) | 
| **Write-up** | **PaperOrchestra** | 
| **Review** | Human peers + autoraters in repo | 

Teams building **[embodied AI criteria benchmarks](https://explainx.ai/blog/embodied-ai-criteria-exposure-benchmark-gaming-pku-2026)** may feed logs directly into `--raw_materials_dir` if they standardize markdown experiment diaries.

## Streamlit demo workflow

The `frontend/` README (linked from the main repo) describes a **Streamlit** interface for non-CLI users. Typical lab workflow:

1. Drop markdown experiment notes into `raw_materials_dir` .
2. Pick a LaTeX template matching your target venue folder.
3. Watch agents stream section drafts (API costs accrue per reflection loop).
4. Download compiled PDF and **diff against human baseline** .

Streamlit is ideal for **PI review**; CLI mode fits **batch rewrites after ablation sweeps**.

## Model provider matrix

| Provider env var | Agent stages typically using it | 
|---|---|
| `GEMINI_API_KEY` / Vertex | Writer + literature (Google-native retrieval tools) | 
| `OPENAI_API_KEY` | Alternative writer or reflection | 
| `SEMANTIC_SCHOLAR_API_KEY` | Citation search rate limits | 

You can assign **cheaper models** to reflection while reserving **frontier models** for related-work synthesis — but cheap reflection may miss **factual errors** in methods sections describing novel architectures like **[MMFlash speculative decoding](https://explainx.ai/blog/mmflash-vision-language-speculative-decoding-november-2026)**.

## How this differs from generic "write my paper" prompts

Single-shot ChatGPT paper requests collapse **structure, citations, and figures** into one context window. PaperOrchestra's separation means:

- Literature agents can **restart retrieval** without wiping methods drafts.
- Plotting agents can **re-execute** matplotlib (or similar) after hyperparameter changes.
- Refinement agents target **section-level** tone instead of whole-document rewrites.

That mirrors production agent design in **[envisioning agentic robotics surveys](https://explainx.ai/blog/envisioning-agentic-robotics-survey-s0-s1-s2-roadmap-2026)** — specialized roles with orchestration, not one omnibus prompt.

## Limits and what to test first

1. **Run without plotting** on a toy template to validate TeX compile on your machine.
2. **Swap writer and reflection models** — cost and quality trade off sharply on long related-work sections.
3. **Check Semantic Scholar rate limits** with your API key tier.
4. **Do not assume CVPR template equals acceptance** — formatting ≠ science.

## Agent roles in more detail

Three grippers place different green shapes into one shallow tray — outline, literature, and reflection agents handing off LaTeX sections

Although the public README groups agents coarsely, the implementation under `methods/` typically separates:

- **Retrieval planning** — what keywords and venues to query (Semantic Scholar API key in`.env` ).
- **Evidence ranking** — which papers enter the related-work section versus appendices.
- **Notation normalization** — keeping symbols consistent across sections (failure mode in single-thread LLM writes).
- **Reflection** — a second model critiques drafts (`--reflection_model_name` ).

Reflection agents mirror **[Claude team code review](https://explainx.ai/blog/claude-code-team-code-review-process-ai-reviewer-risk-tiers-reverts-down-40-percent-2026)** patterns: a different model family can catch hedging or missing limitations paragraphs.

## Templates and conference compliance

`templates/cvpr2025` ships as an example; you can clone the folder for **NeurIPS, ICML, or internal tech reports**. PaperOrchestra does **not** verify you uploaded the latest official style files — maintainers must refresh `.sty` drops themselves.

Before submission:

1. Compile with your local TeXLive.
2. Run venue **PDF tests** (page limit, anonymization).
3. Manually inspect **bibliography keys** — agents love duplicate cites.

## Autoraters and evaluation gap

The `autoraters/` directory hints at **automatic quality scoring**, but without the withheld `datasets/` folder you cannot reproduce Google Cloud AI Research numbers locally. Treat autoraters as **optional dev tooling**, not peer review.

When comparing to **[VibePrevisBench](https://explainx.ai/blog/vibeprevisbench-kling-blender-agent-4d-scene-benchmark-2026)** or gaming-agent papers, decide whether automated writing helps **camera-ready prose** or only **internal tech reports** — many labs ban the former.

## Relationship to Google's other agent releases

PaperOrchestra complements **[Google Stitch Skills](https://explainx.ai/blog/google-stitch-skills-agent-skills-mcp-claude-code-cursor-install-2026)** (coding agents) and **Gemini Robotics** coverage elsewhere on explainx.ai — same company, different surface. None of these repos replace **Vertex governance policies** on customer data.

## Responsible use checklist

1. **Human authorship** — list who verified claims and ran experiments.
2. **Citation spot checks** — open every DOI the literature agent inserted.
3. **Figure provenance** — plotting agents execute code; archive notebooks.
4. **Dual submission policy** — automated variants still count as prior work if posted online.
5. **Export controls** — unrelated to LaTeX, but relevant if your logs include**restricted benchmark numbers** from defense-linked evals.

## Future dataset release

When Google ships the promised `datasets/` drop, rerun PaperOrchestra on their benchmark and compare to your in-house logs — until then, star the repo for **architecture reference**, not **leaderboard scores**.

## Security and secrets handling

The `.env` file holds provider keys and optional SMTP credentials for demo email delivery. Treat generated `output_dir` folders like any research artifact:

- Scrub **API keys** from experiment logs before feeding`raw_materials_dir` .
- Redact **patient identifiers** or**customer names** from traces if your lab works on applied AI such as**[Apple health AI studies](https://explainx.ai/blog/apple-health-ai-usage-study-openevidence-2026)** .
- Do not commit LaTeX outputs with **embedded gen-AI watermarks** to arXiv without venue policy review.

PaperOrchestra is **not** a confidential compute enclave — calls go to configured cloud APIs unless you wire local models yourself (not documented as first-class in the README).

## Star velocity and why researchers noticed

The repo's **+306 stars in a day** spike tracks renewed interest in **agentic science workflows** after **[Edward Witten's Claude-assisted lattice note](https://explainx.ai/blog/edward-witten-claude-opus-5-5-lattice-chiral-fermions-arxiv-2026)** and **[OpenAI lattice science blogging](https://explainx.ai/blog/openai-gpt-6-astra-standard-model-lattice-nathaniel-craig-science-blog-2026)**. Teams want pipelines between **experiment folder** and **arXiv source zip** without hiring a dedicated technical writer for every ablation grid.

Expect forks that add **internal citation styles** (IEEE, company tech report templates) under `templates/` — upstream encourages custom subfolders.

## Integrating PaperOrchestra into a real lab calendar

A pragmatic rollout keeps humans in the loop at three gates: **outline approval**, **related-work spot check**, and **figure recompilation** after final ablations. Schedule the CLI overnight when API rate limits are looser, then let staff edit LaTeX manually the next morning.

Pair with internal git: store `output_dir` as an artifact, tag the commit hash of code used for results, and block auto-submission hooks from pushing to arXiv until a PI signs off — the same discipline you would apply to **[Sakana-style multi-agent training logs](https://explainx.ai/blog/sakana-mass-multi-agent-self-supervision-recursive-self-improvement-2026)**.

*Repository structure and arXiv id match google-research/paper-orchestra README, November 16, 2026.*
