[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-96068":3},{"id":4,"name":5,"fullName":6,"owner":7,"repo":5,"description":8,"homepage":9,"htmlUrl":10,"language":11,"languages":9,"totalLinesOfCode":9,"stars":12,"forks":13,"watchers":14,"openIssues":14,"contributorsCount":9,"subscribersCount":14,"size":14,"stars1d":14,"stars7d":14,"stars30d":14,"stars90d":14,"forks30d":14,"starsTrendScore":14,"compositeScore":15,"rankGlobal":9,"rankLanguage":9,"license":9,"archived":16,"fork":16,"defaultBranch":17,"hasWiki":16,"hasPages":16,"topics":9,"createdAt":9,"pushedAt":9,"updatedAt":18,"readmeContent":19,"aiSummary":20,"trendingCount":14,"starSnapshotCount":14,"syncStatus":21,"lastSyncTime":9,"discoverSource":22},96068,"feynman","advaitpaliwal\u002Ffeynman","advaitpaliwal","The open source AI research agent.",null,"https:\u002F\u002Fgithub.com\u002Fadvaitpaliwal\u002Ffeynman","TypeScript",9168,1048,0,62.06,false,"main","2026-09-20 04:01:32","\u003Cp align=\"center\">\n  \u003Ca href=\"https:\u002F\u002Ffeynman.is\">\n    \u003Cimg src=\"assets\u002Fhero.png\" alt=\"Feynman CLI\" width=\"800\" \u002F>\n  \u003C\u002Fa>\n\u003C\u002Fp>\n\u003Cp align=\"center\">The open source AI research agent.\u003C\u002Fp>\n\u003Cp align=\"center\">\n  \u003Ca href=\"https:\u002F\u002Ffeynman.is\u002Fdocs\">\u003Cimg alt=\"Docs\" src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Fdocs-feynman.is-0d9668?style=flat-square\" \u002F>\u003C\u002Fa>\n  \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fadvaitpaliwal\u002Ffeynman\u002Fblob\u002Fmain\u002FLICENSE\">\u003Cimg alt=\"License\" src=\"https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Flicense\u002Fadvaitpaliwal\u002Ffeynman?style=flat-square\" \u002F>\u003C\u002Fa>\n\u003C\u002Fp>\n\n---\n\n### Installation\n\n**macOS \u002F Linux:**\n\n```bash\ncurl -fsSL https:\u002F\u002Ffeynman.is\u002Finstall | bash\n```\n\n**Windows (PowerShell):**\n\n```powershell\nirm https:\u002F\u002Ffeynman.is\u002Finstall.ps1 | iex\n```\n\nThe one-line installer fetches the latest tagged release. To pin a version, pass it explicitly, for example `curl -fsSL https:\u002F\u002Ffeynman.is\u002Finstall | bash -s -- 0.2.35`.\n\nThe installer downloads a standalone native bundle with its own pinned Node.js runtime and verifies the release SHA-256 before replacing an existing installation.\n\nTo upgrade the standalone app later, rerun the installer. `feynman update` only refreshes installed Pi packages inside Feynman's environment; it does not replace the standalone runtime bundle itself.\n\nTo uninstall the standalone app, remove the launcher and runtime bundle, then optionally remove `~\u002F.feynman` if you also want to delete settings, workbench app state, sessions, and installed package state. If you also want to delete alphaXiv login state, remove `~\u002F.ahub`. See the installation guide for platform-specific paths.\n\n**npm alternative** (uses your local Node.js runtime):\n\n```bash\nnpm install -g @advaitpaliwal\u002Ffeynman\n```\n\nTo update an npm installation, run `npm install -g @advaitpaliwal\u002Ffeynman@latest`.\n\nIf you installed the old scoped package, migrate once:\n\n```bash\nnpm uninstall -g @companion-ai\u002Ffeynman\nnpm install -g @advaitpaliwal\u002Ffeynman\n```\n\nThe command remains `feynman`; the native install commands above are unchanged. See the [installation guide](https:\u002F\u002Ffeynman.is\u002Fdocs\u002Fgetting-started\u002Finstallation) for Node.js requirements and uninstall instructions.\n\nLocal models are supported through the setup flow. For LM Studio, run `feynman setup`, choose `LM Studio`, and keep the default `http:\u002F\u002Flocalhost:1234\u002Fv1` unless you changed the server port. For LiteLLM, choose `LiteLLM Proxy` and keep the default `http:\u002F\u002Flocalhost:4000\u002Fv1`. For Ollama or vLLM, choose `Custom provider (baseUrl + API key)`, use `openai-completions`, and point it at the local `\u002Fv1` endpoint.\n\nTo authenticate another hosted provider, run `feynman model login \u003Cprovider>`. GitHub Copilot sign-in retries model discovery once when GitHub rate-limits the request. OpenRouter login opens an OAuth page and listens for a local callback; over SSH or in another headless environment, paste the browser's final redirect URL or authorization code into Feynman's prompt, or set `OPENROUTER_API_KEY` before launch to use API-key authentication without OAuth.\n\n### Skills Only\n\nIf you want just the research skills without the full terminal app:\n\n**macOS \u002F Linux:**\n\n```bash\ncurl -fsSL https:\u002F\u002Ffeynman.is\u002Finstall-skills | bash\n```\n\n**Windows (PowerShell):**\n\n```powershell\nirm https:\u002F\u002Ffeynman.is\u002Finstall-skills.ps1 | iex\n```\n\nThat installs the skill library into `~\u002F.codex\u002Fskills\u002Ffeynman` for Codex. You can also name the Codex target explicitly:\n\n**macOS \u002F Linux:**\n\n```bash\ncurl -fsSL https:\u002F\u002Ffeynman.is\u002Finstall-skills | bash -s -- --codex\n```\n\n**Windows (PowerShell):**\n\n```powershell\n& ([scriptblock]::Create((irm https:\u002F\u002Ffeynman.is\u002Finstall-skills.ps1))) -Scope Codex\n```\n\nFor a repo-local Claude\u002Fagent install instead:\n\n**macOS \u002F Linux:**\n\n```bash\ncurl -fsSL https:\u002F\u002Ffeynman.is\u002Finstall-skills | bash -s -- --repo\n```\n\n**Windows (PowerShell):**\n\n```powershell\n& ([scriptblock]::Create((irm https:\u002F\u002Ffeynman.is\u002Finstall-skills.ps1))) -Scope Repo\n```\n\nThat installs into `.agents\u002Fskills\u002Ffeynman` under the current repository.\n\nFor an OpenCode project-local install instead:\n\n**macOS \u002F Linux:**\n\n```bash\ncurl -fsSL https:\u002F\u002Ffeynman.is\u002Finstall-skills | bash -s -- --opencode\n```\n\n**Windows (PowerShell):**\n\n```powershell\n& ([scriptblock]::Create((irm https:\u002F\u002Ffeynman.is\u002Finstall-skills.ps1))) -Scope OpenCode\n```\n\nThat installs into `.opencode\u002Fskills\u002Ffeynman` under the current repository.\n\nThese installers download the bundled `skills\u002F` and `prompts\u002F` trees plus the repo guidance files referenced by those skills. They do not install the Feynman terminal, bundled Node runtime, auth storage, or Pi packages.\n\n---\n\n### What you type → what happens\n\n```\n$ feynman \"what do we know about scaling laws\"\n→ Searches papers and web, produces a cited research brief\n\n$ feynman -- \"- summarize the strongest evidence first\"\n→ Preserves a research prompt that begins with a dash instead of parsing it as a CLI option\n\n$ feynman --prompt=\"- summarize the strongest evidence first\"\n→ Runs a dash-leading research prompt once and exits\n\n$ feynman deepresearch \"mechanistic interpretability\"\n→ Multi-agent investigation with parallel researchers, synthesis, verification\n\n$ feynman lit \"RLHF alternatives\"\n→ Literature review with consensus, disagreements, open questions, and lab\u002FPI corpus mode when the input names a research group\n\n$ feynman rank \"mechanistic interpretability sparse autoencoders\"\n→ Decides what to read first with citation, method, reproducibility, and provenance evidence\n\n$ feynman rank \"mechanistic interpretability sparse autoencoders\" --expand-citations 2\n→ Adds cited and citing papers to the local citation graph before scoring graph prestige\n\n$ feynman rank \"mechanistic interpretability sparse autoencoders\" --full-text-top 3\n→ Adds section-aware full-text evidence and checklist rubric answers before rescoring\n\n$ feynman rank \"mechanistic interpretability sparse autoencoders\" --critique-top 5\n→ Adds research-critique strengths, concerns, and follow-up questions grounded in score evidence\n\n$ feynman rank \"mechanistic interpretability sparse autoencoders\" --synthesize\n→ Writes an auditable model synthesis and names the selected model plus whether it was recommended or explicitly requested\n\n$ feynman paper 10.7717\u002Fpeerj.4375 --fetch-full-text\n→ Resolves legal full-text access candidates for one paper and fetches source-specific text when available\n\n$ feynman serve\n→ Opens the standalone science workbench with projects, Pi chat, Feynman Bio Tools, notebooks, compute, artifact previews, provenance, settings, skills, and onboarding context\n\n$ feynman serve --no-auth\n→ Opens the same local workbench at a plain localhost URL for trusted local testing\n\n$ feynman audit 2401.12345\n→ Compares paper claims against the public codebase\n\n$ feynman replicate \"chain-of-thought improves math\"\n→ Plans replication checks and runs them only after an explicit environment choice\n\n$ feynman recipe \"fine-tune a small model for math reasoning\"\n→ Finds ranked, implementable ML training recipes from papers, datasets, docs, and code\n```\n\n---\n\n### Workflows\n\nAsk naturally or use slash commands as shortcuts.\n\n| Command | What it does |\n| --- | --- |\n| `feynman rank \u003Ctopic>` | PaperRank scoring for deciding what to read first, with transparent evidence for citations, methods, reproducibility, and provenance |\n| `feynman paper \u003Cid-or-title>` | Paper access resolver for one DOI, arXiv ID, OpenAlex ID, PMID, PMCID, or title, with OpenAlex, arXiv\u002FalphaXiv, DOI, and Europe PMC candidates plus optional source-specific text fetching |\n| `feynman serve` | Standalone science workbench with project\u002Fsession navigation, project metadata, in-app Pi chat, optional `--no-auth` plain localhost mode, Feynman Bio Tools, Ketcher chemistry sketch artifacts, notebooks, compute, Files host inventory for local, SSH\u002FBYOC, and cloud-backed artifact contexts, audio\u002Fvideo\u002Fspreadsheet\u002Fnotebook\u002FLaTeX\u002Fscience artifact previews including element-level HTML report annotations and KET\u002FRXN\u002FCDXML\u002FCXSMILES chemistry sketches, artifact Notes and note preview modals, Cloud storage credential modal, Cloud export target\u002Fdestination modal, frame records, frame message rows, frame backfill health records, lineage, provenance, settings, org-scoped app-data workbench state under `~\u002F.feynman\u002Forgs\u002F\u003Corg_uuid>\u002Fworkbench`, an org-level `feynman-workbench.db` mirror with physical tables for the full reference-shaped workbench ledger coverage map including compute egress\u002FModal environment fields and Feynman-owned connector ledgers, watch routine state, skill source\u002Flicense state, setup decision state, review feedback state, compute poller lease state, redacted credential state, onboarding intent context, verification files, and `CHANGELOG.md` lab-notebook entries |\n| `\u002Fdeepresearch \u003Ctopic>` | Source-heavy multi-agent investigation |\n| `\u002Flit \u003Ctopic-or-lab>` | Literature review from paper search and primary sources; lab\u002FPI inputs map publication trajectories and originality-ranked papers |\n| `\u002Freview \u003Cartifact>` | Research review with severity and revision plan |\n| `\u002Faudit \u003Citem>` | Paper vs. codebase mismatch audit |\n| `\u002Freplicate \u003Cpaper>` | Plan replication checks; execute only after choosing an environment |\n| `\u002Frecipe \u003Ctask-or-paper>` | Ranked ML training recipes with dataset, method, code, and verification status |\n| `\u002Fcompare \u003Ctopic>` | Source comparison matrix |\n| `\u002Fdraft \u003Ctopic>` | Paper-style draft from research findings |\n| `\u002Fautoresearch \u003Cidea>` | Bounded experiment loop with benchmark evidence |\n| `\u002Fwatch \u003Ctopic>` | Research watch baseline with optional scheduled follow-up |\n| `\u002Fbtw \u003Cquestion>` | Side conversation while the main research agent is busy, with optional handoff back into the main thread |\n| `\u002Fthinking [level]` | View or set model reasoning effort (`off` through `max`, model permitting) without leaving the REPL |\n| `\u002Foutputs` | Browse all research artifacts |\n\n---\n\n### Agents\n\nFour bundled research agents, invoked by workflow prompts when decomposition helps.\n\n- **Researcher** — gather evidence across papers, web, repos, docs\n- **Reviewer** — internal research critique with severity-graded feedback\n- **Writer** — structured drafts from research notes\n- **Verifier** — inline citations, source URL verification, dead link cleanup\n\n---\n\n### Skills & Tools\n\n- **[AlphaXiv](https:\u002F\u002Fwww.alphaxiv.org\u002F)** — paper search, Q&A, code reading, annotations (via Feynman's `alpha` tools and `feynman alpha` command)\n- **Feynman Bio Tools** — Feynman-owned open science connector catalog for literature, exact OpenAlex work\u002Fcitation\u002Freference\u002Fauthor\u002Fvenue workflows, exact arXiv search and batch-paper retrieval, PubMed metadata, PMID\u002FPMCID\u002FDOI conversion, related-article links, citation matching, copyright\u002Flicense checks, PMC full-text routing, bioRxiv\u002FmedRxiv preprint DOI lookup, date\u002Fcategory windows, published-preprint links, funder\u002FROR lookup, usage\u002Fcontent statistics, Europe PMC open-access full-text sections, citation graphs, authors, venues, OA status, ClinicalTrials.gov trial search, NCT details, sponsor programs, eligibility filters, investigator records, endpoint summaries, Grants.gov exact opportunity search, FDA labels, adverse events, recalls, Drugs@FDA applications, sponsor\u002Fstatus\u002Froute counts, pharmacologic classes, generic-equivalent active-ingredient sets, ChEMBL compound\u002Fdrug\u002FADMET\u002Fbioactivity\u002Fmechanism\u002Ftarget workflows, PubChem compound\u002Fsearch\u002Fsimilarity\u002Fbioassay\u002Fsafety workflows, ChEBI entity\u002Fontology workflows, BindingDB target\u002Fcompound workflows, editable Ketcher chemistry sketch seeds, gene, BioMart, Ensembl lookup\u002Fxref\u002FVEP\u002Fhomology\u002Fsequence\u002Foverlap workflows, MyGene query-many, OLS ontology, QuickGO annotation, UniProt entry, Reactome pathway, CellGuide, PanglaoDB marker genes and gene-to-cell-type workflows, exact Antibody Registry antibody\u002FRRID\u002Fcatalog\u002Fstat workflows, reagent, cell-type, metabolomics, genome-track, UCSC exact track\u002Fchromosome\u002Fconservation\u002FTFBS workflows, UniBind TF-DNA binding, KEGG entry\u002Fsearch\u002Flink\u002FID-conversion workflows, InterPro\u002FPfam exact domain architecture, entry, clan, family protein\u002Fproteome modes, Human Protein Atlas exact gene\u002Fsearch modes, STRING exact ID mapping, network, similarity, and best-hit workflows, purchasable ZINC compounds, protein, predicted-structure, structure, EM-map, complex, interaction, exact ENCODE\u002FJASPAR\u002FUniBind regulation modes for experiments, biosamples, files, matrices, species\u002Ftaxa\u002Fcollections\u002Freleases, datasets, and regional TFBS, exact ArrayExpress\u002FGEO\u002FMetaboLights\u002FMGnify\u002FPRIDE omics-archive modes for experiments, samples, files, analyses, projects, and protein evidence, metagenomics, chemical-ontology, chemistry, pathway, exact Rfam RNA family metadata, accession\u002Fid conversion, seed alignment, covariance model, tree, region, structure-mapping, and sequence-search workflows, exact gnomAD short variant, gene, region, liftover, ClinVar-mirror, structural, and mitochondrial workflows, exact CADD variant\u002Fposition\u002Frange scores, exact direct ClinVar search\u002Faccession\u002FrsID workflows, exact dbSNP rsID\u002Fregion workflows, GWAS Catalog exact association\u002Fstudy\u002Ftrait\u002FSNP workflows, eQTL Catalogue exact dataset and association workflows, PheWeb\u002FFinnGen PheWAS workflows, GTEx dataset\u002Ftissue\u002Fsample\u002Fgene\u002Fexpression\u002FeQTL workflows, tissue\u002Fprotein-atlas, expression, human-genetics, cBioPortal study\u002Fdetail\u002Fmutation-frequency\u002Fmutation\u002FCNA\u002Fclinical-attribute workflows, DepMap model\u002Fgene\u002Fdependency workflows, CIViC gene\u002Fvariant\u002Fevidence\u002Fassertion\u002Fprofile\u002Fdisease\u002Ftherapy workflows, ClinGen validity\u002Fdosage\u002Factionability\u002Fclassification workflows, Open Targets disease-drug\u002Fdisease-target\u002Fdrug\u002Fsearch workflows, and canceromics sources\n- **[Hugging Face Hub](https:\u002F\u002Fhuggingface.co\u002Fdocs\u002Fhub\u002Fapi)** — dataset metadata, split\u002Fschema inspection, and small file reads from model, dataset, and Space repos\n- **Web research** — multi-provider search, explicit proxy routing, bounded GitHub issue\u002FPR documents, raw or question-grounded page retrieval, direct images, external fetched-content caching, stored-page passage lookup, and auditable source text; tools, commands, images, PDFs, and browser cookies remain independently gated\n- **Session search** — indexed recall across prior research sessions\n- **Artifact previews** — local workbench viewers for reports, JSON\u002FJSONL, tables, PDFs, images, audio, video, XLSX workbooks, Jupyter notebooks, LaTeX, sequences, alignments, variants, genomes, KET\u002FRXN\u002FCDXML\u002FCXSMILES\u002FMolfile\u002FSDF\u002FSMILES chemistry artifacts, structures, trees, and tensors\n- **Observability** — PostHog analytics, logs, distributed traces, and Pi AI runtime traces through OpenTelemetry metadata, with signal-specific HTTP OTLP routing for external collectors\n- **Research execution options** — Docker, Modal, and RunPod instructions for explicitly chosen replication, benchmark, or dataset-heavy experiment runs; not service deployment or generic cloud administration\n- **Workbench control plane** — local onboarding, project\u002Fsession\u002Fframe state, upload-frame linkage, frame message rows, frame backfill health records, chat-produced artifact attachment, artifact\u002Fversion lineage, Files host inventory for local workspace files, SSH\u002FBYOC compute hosts, and cloud buckets, media\u002Fdocument\u002Fscience previews, element-level HTML report annotations, artifact Notes and note preview modals, Cloud storage credential modal, Cloud export target\u002Fdestination modal with audit logs, execution logs, verification checks, memory categories, watch routine records, skill source\u002Flicense records, setup decision records, review feedback records, compute poller lease records, scoped settings, and redacted credential availability records under Feynman's own runtime and workspace files\n\n---\n\n### How it works\n\nBuilt on [Pi](https:\u002F\u002Fgithub.com\u002Fbadlogic\u002Fpi-mono) for the agent runtime, [alphaXiv](https:\u002F\u002Fwww.alphaxiv.org\u002F) for paper search and analysis, and CLI tools for compute and execution. Runtime resources follow Pi's documented package model for [packages](https:\u002F\u002Fgithub.com\u002Fbadlogic\u002Fpi-mono\u002Fblob\u002Fmain\u002Fpackages\u002Fcoding-agent\u002Fdocs\u002Fpackages.md), [extensions](https:\u002F\u002Fgithub.com\u002Fbadlogic\u002Fpi-mono\u002Fblob\u002Fmain\u002Fpackages\u002Fcoding-agent\u002Fdocs\u002Fextensions.md), and [skills](https:\u002F\u002Fgithub.com\u002Fbadlogic\u002Fpi-mono\u002Fblob\u002Fmain\u002Fpackages\u002Fcoding-agent\u002Fdocs\u002Fskills.md). Hugging Face inspection uses the public [Hub API endpoints](https:\u002F\u002Fhuggingface.co\u002Fdocs\u002Fhub\u002Fapi) and `HF_TOKEN` \u002F `HUGGINGFACE_HUB_TOKEN` environment variables documented by [`huggingface_hub`](https:\u002F\u002Fhuggingface.co\u002Fdocs\u002Fhuggingface_hub\u002Fmain\u002Fen\u002Fpackage_reference\u002Fenvironment_variables). The ML recipe workflow was informed by the open-source [Hugging Face `ml-intern`](https:\u002F\u002Fgithub.com\u002Fhuggingface\u002Fml-intern) research-agent repo, but is implemented as native Feynman prompts, skills, and read-only tools. Research outputs are source-grounded — research claims link to papers, docs, or repos with direct URLs.\n\n---\n\n### Star History\n\n\u003Ca href=\"https:\u002F\u002Fwww.star-history.com\u002F?repos=advaitpaliwal%2Ffeynman&type=date&legend=top-left\">\n  \u003Cpicture>\n    \u003Csource media=\"(prefers-color-scheme: dark)\" srcset=\"https:\u002F\u002Fapi.star-history.com\u002Fchart?repos=advaitpaliwal\u002Ffeynman&type=date&theme=dark&legend=top-left\" \u002F>\n    \u003Csource media=\"(prefers-color-scheme: light)\" srcset=\"https:\u002F\u002Fapi.star-history.com\u002Fchart?repos=advaitpaliwal\u002Ffeynman&type=date&legend=top-left\" \u002F>\n    \u003Cimg alt=\"Star History Chart\" src=\"https:\u002F\u002Fapi.star-history.com\u002Fchart?repos=advaitpaliwal\u002Ffeynman&type=date&legend=top-left\" \u002F>\n  \u003C\u002Fpicture>\n\u003C\u002Fa>\n\n---\n\nThe bundled research runtime is updated as a coordinated set, including Pi, Alpha Hub's `alpha-mcp`, document parsing, web research, and subagents. See the [package stack](https:\u002F\u002Ffeynman.is\u002Fdocs\u002Freference\u002Fpackage-stack) and [release notes](https:\u002F\u002Ffeynman.is\u002Fdocs\u002Freference\u002Freleases) for versions and upgrade details.\n\n### Contributing\n\nSee [CONTRIBUTING.md](CONTRIBUTING.md) for the full contributor guide.\n\n```bash\ngit clone https:\u002F\u002Fgithub.com\u002Fadvaitpaliwal\u002Ffeynman.git\ncd feynman\nnvm use || nvm install\nnpm install\nnpm test\nnpm run typecheck\nnpm run build\n```\n\n[Docs](https:\u002F\u002Ffeynman.is\u002Fdocs) · [Release Notes](RELEASES.md) · [MIT License](LICENSE)\n","Feynman 是一个开源的 AI 研究智能体，专为科研人员与技术工作者设计，支持本地与云端大模型接入、多步推理、文献检索（如 arXiv）、代码生成与执行、数学推导及结构化知识整理。其核心特点是轻量级 CLI 工具、内置独立 Node.js 运行时、零依赖一键安装，并原生支持 LM Studio、Ollama、LiteLLM、OpenRouter 及 GitHub Copilot 等多种模型后端。适用于学术研究辅助、技术文档分析、论文阅读理解、实验方案设计等需要深度 AI 协作的科研场景。",2,"trending"]