doiget
Converts DOIs and arXiv IDs into local PDFs via official open-access APIs, designed as an agent-facing companion for academic paper retrieval.
README
doiget
A single-binary CLI + stdio MCP server that turns DOIs and arXiv ids into local PDFs through official, OA-first APIs. Designed as the agent-facing companion to BiblioFetch.jl.
Docs: stable = the Zola site (built from main); dev = rustdoc built from next; API = docs.rs (latest published release).
Status: Shipping on crates.io (doiget-core, doiget-cli,
doiget-mcp), with sigstore-signed binaries + an SBOM attached to the GitHub
Release. Tier 1 + Tier 2 sources, the stdio MCP server, citation-graph
expansion, and gated TDM sources are all implemented. Releases are cut by a
single signed git tag through the tag-driven pipeline (see
ADR-0025); release-plz was
retired. See CHANGELOG.md for history and
docs/PHASES.md for the phase plan.
Posture
doiget is a general-purpose automation tool for retrieving papers via official publisher APIs. By design, doiget only attempts retrieval through:
- Public Open Access sources (default — Crossref, Unpaywall, arXiv).
- Credentials the user has personally configured for their own institutional or personal subscriptions (opt-in, compile-time gated).
doiget does not work around any access control mechanism, redistribute papers, host content, operate as a SaaS, or bundle any publisher API keys.
Users are responsible for ensuring they have the right to access the content they request and for compliance with each source's Terms of Service.
See docs/LEGAL.md and docs/SCOPE.md.
Documentation
| Reader | Entry point |
|---|---|
| CLI user | This README, then doiget --help, then docs/CONFIG.md and docs/ERRORS.md for non-trivial flags / exit codes |
| Agent / MCP host integrator | docs/MCP_TOOLS.md + docs/INTEGRATION/README.md |
| Library user (Rust) | docs/PUBLIC_API.md + crates.io rustdoc |
| Contributor | CONTRIBUTING.md → docs/ARCHITECTURE.md → docs/DECISIONS/ |
| Publisher legal team | docs/LEGAL.md + CONTACT.md |
| Security researcher | docs/SECURITY.md + docs/PROVENANCE_LOG.md + docs/CAPABILITY.md |
| BiblioFetch.jl user | docs/MIGRATION.md + docs/STORE.md + docs/SAFEKEY.md |
Architecture: docs/ARCHITECTURE.md Threat model: docs/SECURITY.md Permanent non-goals: docs/SCOPE.md Phase plan: docs/PHASES.md ADRs: docs/DECISIONS/
Installation
doiget ships a single self-contained binary — the Linux build is fully static (musl), so it runs on old glibc / HPC boxes too. Every channel installs the same checksum-verified binary from the signed GitHub Release.
Shell installer (Linux / macOS)
curl -fsSL https://raw.githubusercontent.com/sotashimozono/doiget/main/scripts/install.sh | sh
Installs to ~/.local/bin (override with DOIGET_INSTALL_DIR); pin a version with
DOIGET_VERSION=0.6.0. The script verifies the published SHA-256 sidecar before installing.
PowerShell installer (Windows)
irm https://raw.githubusercontent.com/sotashimozono/doiget/main/scripts/install.ps1 | iex
From crates.io (Rust toolchain)
cargo install doiget-cli # installs the `doiget` binary
Further prebuilt-binary channels (Homebrew tap, cargo binstall, npm/npx, Nix flake,
.deb) are tracked in #247. Every
release asset is cosign-keyless signed (<asset>.cosign.bundle) for optional verification.
Quick start
# Fetch a paper by DOI (Open Access only by default)
doiget fetch 10.1103/PhysRevLett.130.200601
# Fetch by arXiv id
doiget fetch arXiv:2401.12345
# Batch fetch
doiget batch refs.txt
# Verify a bibliography's references resolve (no PDF download) — CI gate
doiget verify docs/references.bib --strict
# Lint a .bib for structural issues (no network): missing fields,
# blank fields, $$-display-math titles. Read-only and math-aware.
doiget lint docs/references.bib
# Inspect what was fetched
doiget info 10.1103/PhysRevLett.130.200601
# Run as MCP server (stdio)
doiget serve
Default features fetch only Open Access PDFs through Crossref / Unpaywall / arXiv. Institutional TDM access (Springer OA, APS Harvest, Elsevier ScienceDirect TDM) is not in the default published binary; it must be opted in at build time per publisher. See docs/SOURCES.md.
Coexistence with BiblioFetch.jl
doiget and BiblioFetch.jl share the same on-disk store format (TOML metadata + PDF files
under a configurable store root). doiget defaults to ./papers (under the current working
directory; ADR-0036), BiblioFetch.jl to ~/papers/; point both at the same root (e.g.
DOIGET_STORE_ROOT=~/papers) to share one store. The shared schema, locking protocol, and
atomic write contract are
specified in docs/STORE.md. Reference test vectors for the shared safekey
algorithm are in docs/SAFEKEY.md.
| Tool | Best for |
|---|---|
| BiblioFetch.jl | Julia REPL, research vault, citation graph exploration |
| doiget | Agents / MCP hosts, batch operations, scripted pipelines, container deployments |
License
MIT for the doiget source code and binaries (see LICENSE).
The license under which doiget retrieves papers is separate and is determined by each paper's own license, the publisher's API Terms of Service, and the user's own access rights. doiget does not relicense fetched content.
Contact
Takedown requests, security disclosures, and other formal correspondence: CONTACT.md.
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。