[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-80107":3},{"id":4,"name":5,"fullName":6,"owner":7,"repo":5,"description":8,"homepage":9,"htmlUrl":10,"language":11,"languages":10,"totalLinesOfCode":10,"stars":12,"forks":13,"watchers":14,"openIssues":15,"contributorsCount":13,"subscribersCount":13,"size":13,"stars1d":13,"stars7d":13,"stars30d":16,"stars90d":13,"forks30d":13,"starsTrendScore":13,"compositeScore":17,"rankGlobal":10,"rankLanguage":10,"license":18,"archived":19,"fork":19,"defaultBranch":20,"hasWiki":21,"hasPages":21,"topics":22,"createdAt":10,"pushedAt":10,"updatedAt":30,"readmeContent":31,"aiSummary":32,"trendingCount":13,"starSnapshotCount":13,"syncStatus":15,"lastSyncTime":33,"discoverSource":34},80107,"Plato-Scientific-Research-Autonomous-Agent","Eldergenix\u002FPlato-Scientific-Research-Autonomous-Agent","Eldergenix","Multi-agent AI scientist that turns experimental data into > publication-ready research papers.","",null,"Python",53,0,56,2,1,40.1,"GNU General Public License v3.0",false,"main",true,[23,24,25,26,27,28,29],"agent-framework","agentic-science","agentic-workflow","ai-agents","ai-research","ai-tools","science-research","2026-06-12 04:01:26","# Plato\n\n[![Python Version](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Fpython-%3E%3D3.12-blue.svg)](https:\u002F\u002Fwww.python.org\u002Fdownloads\u002F) [![License: GPL v3](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FLicense-GPLv3-blue.svg)](https:\u002F\u002Fwww.gnu.org\u002Flicenses\u002Fgpl-3.0)\n\nPlato is a multi-agent AI system that takes experimental data and produces peer-reviewable scientific papers end-to-end — generating the research idea, designing the methodology, running the analysis, and writing the LaTeX manuscript through a reviewer-panel revision loop.\n\n## What's new in 0.2\n\nPhase 5 hardening landed alongside the dashboard's 13-stream feature push:\n\n- **Multi-source retrieval** — six adapters (arXiv, OpenAlex, ADS, Crossref,\n  PubMed, Semantic Scholar) behind a domain-aware orchestrator with rate-limit\n  backoff, ETag caching, and per-host circuit breakers.\n- **Citation validation** — every reference is resolved against Crossref +\n  Retraction Watch + arXiv before the paper finalizes. The run dir gets a\n  `validation_report.json` with per-reference pass\u002Ffail.\n- **Claim → Evidence Matrix** — the literature pass extracts atomic claims\n  with quote spans and links them to source records. Persisted as\n  `evidence_matrix.jsonl` per run.\n- **Reviewer panel + revision loop** — methodology \u002F statistics \u002F novelty \u002F\n  writing axes feed an aggregator that drives a redraft loop bounded by\n  `Plato.get_paper(max_revision_iters=...)`.\n- **Autonomous research loop** — `plato loop --hours 8 --max-cost-usd 50`\n  iterates under a wall-clock + cost budget, committing improvements and\n  reverting regressions.\n- **Reproducibility manifest** — every workflow emits `manifest.json` with\n  git sha, project sha-256, model versions, source ids, tokens, and cost.\n- **Observability** — opt in by setting `LANGFUSE_*` env vars; LangFuse\n  callbacks are wired into every LangGraph invocation.\n- **Pluggable domains** — `DomainProfile` registry exposes retrieval,\n  keyword extractor, journal preset, executor, and novelty corpus as swap\n  points. Astro is the default; biology ships out-of-the-box.\n- **Multi-tenant dashboard** — set `PLATO_DASHBOARD_AUTH_REQUIRED=1` and\n  the dashboard reads `X-Plato-User` from the upstream proxy to scope\n  every project, key store, and run artifact per tenant.\n\nSee `docs\u002Fadr\u002F` for the design decisions behind these changes and\n`dashboard\u002FCHANGELOG.md` for the full list.\n\n## Resources\n\n- [Live Plato demo](https:\u002F\u002Fplato-production-9fea.up.railway.app)\n\n- [GitHub repository](https:\u002F\u002Fgithub.com\u002FEldergenix\u002FPlato-Scientific-Research-Autonomous-Agent)\n\n- [Paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2510.26887)\n\n\n## Installation\n\nTo install plato create a virtual environment and pip install it. We recommend using Python 3.12:\n\n```bash\npython -m venv Plato_env\nsource Plato_env\u002Fbin\u002Factivate\npip install \"plato[dashboard]\"\n```\n\nOr alternatively install it with [uv](https:\u002F\u002Fdocs.astral.sh\u002Fuv\u002F), initializing a project and installing it:\n\n```bash\nuv init\nuv add plato[dashboard]\n```\n\nThen, run the Plato dashboard with:\n\n```\nplato dashboard\n```\n\n## Get started\n\nInitialize a `Plato` instance and describe the data and tools to be employed.\n\n```python\nfrom plato import Plato\n\np = Plato(project_dir=\"project_dir\")\n\nprompt = \"\"\"\nAnalyze the experimental data stored in data.csv using sklearn and pandas.\nThis data includes time-series measurements from a particle detector.\n\"\"\"\n\np.set_data_description(prompt)\n```\n\nGenerate a research idea from that data specification.\n\n```python\np.get_idea()\n```\n\nGenerate the methodology required for working on that idea.\n\n```python\np.get_method()\n```\n\nWith the methodology setup, perform the required computations and get the plots and results.\n\n```python\np.get_results()\n```\n\nFinally, generate a latex article with the results. You can specify the journal style, in this example we choose the [APS (Physical Review Journals)](https:\u002F\u002Fjournals.aps.org\u002F) style.\n\n```python\nfrom plato import Journal\n\np.get_paper(journal=Journal.APS)\n```\n\nYou can also manually provide any info as a string or markdown file in an intermediate step, using the `set_idea`, `set_method` or `set_results` methods. For instance, for providing a file with the methodology developed by the user:\n\n```python\np.set_method(path_to_the_method_file.md)\n```\n\n## Plato Dashboard (new, recommended)\n\nA Linear-themed real-time web dashboard with full pipeline visualization, cost tracking, and live agent log streaming. See [dashboard\u002FREADME.md](dashboard\u002FREADME.md) for setup.\n\n```bash\npip install \"plato[dashboard]\"\nplato dashboard\n```\n\nThe dashboard supersedes the legacy `plato run` Streamlit app for new workflows.\n\nFor hosted Railway SaaS\u002FLab deployments with Clerk auth or Clerk Billing, run\nthe local production gates and redacted strict preflight before deploying:\n\n```bash\nbash dashboard\u002Fscripts\u002Fcheck-local-production-gates.sh\nbash dashboard\u002Fscripts\u002Fcheck-hosted-saas-preflight.sh --railway --service plato --environment production --hosted-required --strict\n```\n\nIt verifies the Clerk user\u002FLab auth contract, `PLATO_BACKEND_PROXY_SECRET`,\npublic origin, Clerk proxy, and hosted billing flags without printing secret\nvalues. Strict mode treats preflight warnings as release blockers. After\ndeploying, use the read-only production readiness check to probe the live app\nand scan Railway logs:\n\n```bash\nbash dashboard\u002Fscripts\u002Fcheck-production-readiness.sh --service plato --environment production --origin https:\u002F\u002Fdiscovering.app\n```\n\nIf `railway variables --json`\u002F`--kv` is unavailable but you have a local\nRailway variables snapshot, pass it with `--variables-file`; the checker still\nprints only redacted key status and lengths:\n\n```bash\nbash dashboard\u002Fscripts\u002Fcheck-production-readiness.sh --service plato --environment production --origin https:\u002F\u002Fdiscovering.app --variables-file \u002Fpath\u002Fto\u002Frailway-variables.json\n```\n\nSee [dashboard\u002FRAILWAY.md](dashboard\u002FRAILWAY.md) for the full production\nvariable matrix.\n\n## Build from source\n\n### pip\n\nYou will need python 3.12 or higher installed. Clone Plato:\n\n```bash\ngit clone https:\u002F\u002Fgithub.com\u002FEldergenix\u002FPlato-Scientific-Research-Autonomous-Agent.git\ncd Plato-Scientific-Research-Autonomous-Agent\n```\n\nCreate and activate a virtual environment\n\n```bash\npython3 -m venv Plato_env\nsource Plato_env\u002Fbin\u002Factivate\n```\n\nAnd install the project\n\n```bash\npip install -e .\n```\n\n### uv\n\nYou can also install the project using [uv](https:\u002F\u002Fdocs.astral.sh\u002Fuv\u002F), just running:\n\n```bash\nuv sync\n```\n\nwhich will create the virtual environment and install the dependencies and project. Activate the virtual environment if needed with\n\n```bash\nsource .venv\u002Fbin\u002Factivate\n```\n\n## Docker\n\nYou can run Plato with [Docker](https:\u002F\u002Fwww.docker.com\u002F) using the dashboard compose file:\n\n```bash\ndocker compose -f dashboard\u002Fcompose.yaml up --build\n```\n\nThe local dashboard runs on `http:\u002F\u002Flocalhost:7878` by default.\n\nYou can also build an image locally with\n\n```bash\ndocker build -f docker\u002FDockerfile.dev -t plato_src .\n```\n\n## Contributing\n\nPull requests are welcome! Feel free to open an issue for bugs, comments, questions and suggestions.\n\n\u003C!-- ## Citation\n\nIf you use this library please link this repository and cite [arXiv:2506.xxxxx](arXiv:x2506.xxxxx). -->\n\n## Citation\n\nIf you make use of Plato, please cite the following references:\n\n```bibtex\n@article{villaescusanavarro2025platoprojectdeepknowledge,\n         title={The Plato project: Deep knowledge AI agents for scientific discovery}, \n         author={Francisco Villaescusa-Navarro and Boris Bolliet and Pablo Villanueva-Domingo and Adrian E. Bayer and Aidan Acquah and Chetana Amancharla and Almog Barzilay-Siegal and Pablo Bermejo and Camille Bilodeau and Pablo Cárdenas Ramírez and Miles Cranmer and Urbano L. França and ChangHoon Hahn and Yan-Fei Jiang and Raul Jimenez and Jun-Young Lee and Antonio Lerario and Osman Mamun and Thomas Meier and Anupam A. Ojha and Pavlos Protopapas and Shimanto Roy and David N. Spergel and Pedro Tarancón-Álvarez and Ujjwal Tiwari and Matteo Viel and Digvijay Wadekar and Chi Wang and Bonny Y. Wang and Licong Xu and Yossi Yovel and Shuwen Yue and Wen-Han Zhou and Qiyao Zhu and Jiajun Zou and Íñigo Zubeldia},\n         year={2025},\n         eprint={2510.26887},\n         archivePrefix={arXiv},\n         primaryClass={cs.AI},\n         url={https:\u002F\u002Farxiv.org\u002Fabs\u002F2510.26887},\n}\n\n@software{Plato_2025,\n          author = {Pablo Villanueva-Domingo, Francisco Villaescusa-Navarro, Boris Bolliet},\n          title = {Plato: Modular Multi-Agent System for Scientific Research Assistance},\n          year = {2025},\n          url = {https:\u002F\u002Fgithub.com\u002FEldergenix\u002FPlato-Scientific-Research-Autonomous-Agent},\n          note = {Available at https:\u002F\u002Fgithub.com\u002FEldergenix\u002FPlato-Scientific-Research-Autonomous-Agent},\n          version = {latest}\n          }\n\n@software{CMBAGENT_2025,\n          author = {Boris Bolliet},\n          title = {CMBAGENT: Open-Source Multi-Agent System for Science},\n          year = {2025},\n          url = {https:\u002F\u002Fgithub.com\u002FCMBAgents\u002Fcmbagent},\n          note = {Available at https:\u002F\u002Fgithub.com\u002FCMBAgents\u002Fcmbagent},\n          version = {latest}\n          }\n```\n\n## License\n\n[GNU GENERAL PUBLIC LICENSE (GPLv3)](https:\u002F\u002Fwww.gnu.org\u002Flicenses\u002Fgpl-3.0.html)\n\nPlato - Copyright (C) 2026 Pablo Villanueva-Domingo, Francisco Villaescusa-Navarro, Boris Bolliet\n","Plato 是一个多智能体AI系统，能够将实验数据转化为可供同行评审的科学论文。其核心功能包括从生成研究想法、设计方法论、运行数据分析到撰写LaTeX手稿，并通过审稿人小组的修订循环来完善论文。技术特点上，Plato支持多源检索、引用验证、证据矩阵构建、自动研究循环等高级特性，并且具有高度可观察性和领域可插拔性，确保了研究过程的透明度和灵活性。适用于需要高效处理科研数据并快速产出高质量学术成果的研究场景中。","2026-06-11 03:59:16","CREATED_QUERY"]