convez.ai
MCP ServerBusca & PesquisaIntermediário
cbyrohl

MCP Server NASA ADS

cbyrohl/mcp-server-ads

Acesso ao maior banco de dados de astrofísica via MCP

Ver no GitHub
InicialPluginScore · de 0 a 100

Servidor MCP que integra o NASA Astrophysics Data System (ADS) ao Claude e outras ferramentas compatíveis. Permite buscar artigos astrofísicos, navegar grafos de citações, exportar em múltiplos formatos e gerenciar listas de leitura através de linguagem natural.

Instalação

$npx convez-ai cbyrohl/mcp-server-ads

1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.

Métricas

Estrelas
6
Forks
0
Acompanhando
0
Saúde
42%
Último commit
há 5 meses
Versão
—
Issues abertas
0
Contribuidores
1

Casos de uso

  • Pesquisa bibliográfica em astrofísica com análise de citações
  • Exportação de referências em BibTeX, AASTeX e outros formatos
  • Análise de métricas de citação e tendências de publicação

Veredito

Ferramenta especializada bem construída para pesquisadores em astrofísica. Oferece integração robusta com o ADS via MCP, mas seu valor é restrito ao ecossistema acadêmico dessa disciplina.

README

A tradução para o português deste README ainda está na fila.

mcp-server-ads

Python 3.11+ License: MIT

A powerful MCP server for the NASA Astrophysics Data System (ADS) — the primary database for astrophysics literature. Search papers, traverse citation graphs forward and backward, export BibTeX, compute metrics, and manage reading lists, all through natural language. Works with Claude Desktop/Code, Cursor, OpenAI Codex, and any MCP-compatible client.

demo_15fps_crf20.webm

Quick Start

Get an ADS API Token

  1. Create a free account at NASA ADS
  2. Log in and go to Settings > API Token
  3. Click Generate a new key and copy the token

Installation

No separate install step needed — the client configurations below use uvx to automatically fetch and run the server. Just pick your client and go. Requires Python 3.11+ and uv.

Claude Code

claude mcp add --scope user mcp-server-ads -e ADS_API_TOKEN=your-api-token-here -- uvx --from git+https://github.com/cbyrohl/mcp-server-ads mcp-server-ads

Use --scope project instead to share the configuration via .mcp.json in your repo, or omit --scope for local (current project only).

Claude Desktop

Add to your Claude Desktop config:

{
  "mcpServers": {
    "ads": {
      "command": "uvx",
      "args": ["--from", "git+https://github.com/cbyrohl/mcp-server-ads", "mcp-server-ads"],
      "env": {
        "ADS_API_TOKEN": "your-api-token-here"
      }
    }
  }
}

Codex CLI

codex mcp add mcp-server-ads --env ADS_API_TOKEN=your-api-token-here -- uvx --from git+https://github.com/cbyrohl/mcp-server-ads mcp-server-ads

This installs to ~/.codex/config.toml (user-level, available across all projects). For project-scoped config, add the entry to .codex/config.toml in your project root instead.

Standalone / Python API

If you want to install the package directly (e.g. as a Python library or to run the server manually):

# Install with uv
uv tool install git+https://github.com/cbyrohl/mcp-server-ads

# Or with pip
pip install git+https://github.com/cbyrohl/mcp-server-ads

Running from Source

git clone https://github.com/cbyrohl/mcp-server-ads.git
cd mcp-server-ads
uv sync

# Run the server
ADS_API_TOKEN=your-token uv run mcp-server-ads

Configuration

Environment VariableRequiredDefaultDescription
ADS_API_TOKENYes—API token from ADS
ADS_API_URLNohttps://api.adsabs.harvard.eduAPI base URL (override for SciX)

Tools (11)

Search

ToolDescription
ads_searchSearch the ADS database with full query syntax, including citations(), references(), similar(), trending(), and reviews() operators
ads_bigquerySearch within a specific set of bibcodes (up to 2000)

Export & Metrics

ToolDescription
ads_exportExport records in 18+ formats (BibTeX, AASTeX, RIS, CSL, etc.)
ads_metricsCompute citation metrics (h-index, g-index, citation counts, etc.)

Libraries

ToolDescription
ads_libraryManage libraries: list, get, create, edit, or delete saved paper collections
ads_library_documentsManage documents and notes within a library: add/remove papers, set operations (union, intersection, difference, copy, empty), and note CRUD

Discovery & Resolution

ToolDescription
ads_resolve_linksResolve available links for a paper (full text, data, etc.)
ads_object_searchTranslate astronomical object names to ADS queries (SIMBAD/NED)
ads_citation_helperSuggest papers that should be cited alongside a given set
ads_resolve_referenceResolve free-text reference strings to ADS bibcodes

Network Visualization

ToolDescription
ads_networkGenerate author collaboration or paper citation networks from a set of papers

Resources

URIDescription
ads://fieldsComplete reference of searchable and returnable ADS fields
ads://syntaxADS query syntax quick-reference with examples
ads://rate-limitsLive API rate-limit status

Prompts

PromptDescription
literature_reviewMulti-step literature review workflow for a research topic
citation_analysisCitation network analysis workflow for a set of papers
generate_bibliographyGenerate a formatted bibliography from search or bibcodes

Similar Projects

This server focuses on broad ADS API coverage, token-efficient output, and integration-tested reliability. Compared to arXiv-based tools, ADS enables forward/backward citation traversal (citations and references of any paper), though it is focused on astrophysics.

Development

# Install dev dependencies
uv sync

# Run tests
uv run pytest

# Lint
uv run ruff check src/ tests/

# Run the server locally
ADS_API_TOKEN=your-token uv run mcp-server-ads

License

MIT

Ver o original no GitHub

Bom para

Relacionados