Mrd

io.github.newspacemarket-com由 GitHub 账号 newspacemarket-com 发布v1.2.0更新于 Oct 24, 2026

Spacecraft component specs from vendor datasheets: search, digests, facts, gaps, budgets, bus design

已验证Streamable HTTP可网页运行知识与记忆开发者工具数据与分析

概览

AI 生成的概览

检索并比较来自厂商数据手册的航天器部件规格,支持需求匹配、质量与功率预算以及整星总线设计。

可以做什么
  • 按工程需求反向搜索部件,候选按满足、不满足或数据缺失分组
  • 读取类型化规格、属性定义和数据手册摘要,均附页码出处
  • 跨整个类别查询某一事实字段,或就某份数据手册提问并取得相关摘录
  • 查找某产品的替代件、供应商和数据缺口
  • 为部件组合计算质量与功率预算、检查接口适配,并按自然语言任务组装整星总线
适用场景
适合在挑选或比较反作用轮、星敏感器、无线电、结构件等飞行硬件时使用,也适合检查部件之间能否对接,或生成可追溯来源的质量与功率预算。
运行要求
可作为远程托管端点使用,无需安装;也可通过 npx 运行本地 npm 包,转发到托管服务。读取无需认证;可选的免费 MRD_API_KEY 可将速率限制提高五倍。需要网络访问。
安装前请注意
这是只读数据服务,但目录并不完整:厂商未提供机器可读资料的产品不会收录,缺失字段会如实标注而非填补。数值为机器抽取并带置信度等级,关键数据应核对所引用的数据手册页面。

安装

在 SourceWeft 中

  1. 打开 控制台中的 Mrd,将其添加到工作区。
  2. 为需要使用其工具的对话启用该服务。

Web executable,通过 Streamable HTTP。 远程服务在工作区中配置后即可从网页运行时运行。

交给智能体安装

把这段提示词粘贴到 Claude Code、Codex、Cursor 等智能体,或 SourceWeft 对话中。智能体会读取该服务器的安装指南,先向你展示发布者、端点和登录方式,经你同意后再把它添加到你的客户端。密钥不会写进任何命令。

阅读 https://sourceweft.com/mcp/io.github.newspacemarket-com%2Fmrd/install.md,并按其中说明把这个 MCP 服务器接入我的客户端。修改任何配置之前,先告诉我它的发布者、端点和登录方式,等我确认。不要把 API 密钥写进命令或我会提交的文件;问我密钥应从哪里获取。

查看智能体所依据的安装指南

其他 MCP 客户端

把它添加到你客户端的 mcpServers 配置中。

{
  "mcpServers": {
    "mrd": {
      "type": "http",
      "url": "https://newspacemarket.com/mrd-connect/mcp"
    }
  }
}

README

MRD — machine-readable datasheets for spacecraft components

[MCP Registry] [Smithery] [Data licence] [Code licence]

Typed, datasheet-sourced specifications for 2,500+ flight hardware products across 24 categories — reaction wheels, star trackers, propulsion, radios, antennas, structures, ground stations and more — with 19,900+ values that each cite the datasheet page they came from. Underneath sits the datasheet layer itself: 26,000+ facts read from 800+ datasheets, each anchored to a quoted page, and the full text of 1,600+ datasheets indexed for questions.

Built for software and AI assistants, not only for reading. Ask in plain language, filter by engineering requirements, or download the whole thing.

"Find a reaction wheel over 0.2 N·m·s that runs on a 12 V bus, under 2 kg"
A  meets every requirement          VRW-B-02 · Vectronic · 0.2 N·m·s · 9–36 V · 1000 gB  fails a named requirement        RW-0.06 · momentum 0.06 N·m·s — need ≥ 0.2C  could fit, data not published    NanoTorque GSW-600 · missing: supply voltage

Three rules make the data usable by a machine:

  • Nothing is guessed. A missing property means the vendor does not publish it, and the item is reported as "data missing" rather than silently dropped or filled with a typical value.
  • Conditions are kept apart. Idle, steady-state and peak power are separate records with their measurement condition attached — never averaged into one number.
  • Every value is traceable. Source file and page, plus a confidence level: L0 machine-extracted, L1 verified by a human against the vendor datasheet.

Use it from an AI assistant (MCP)

Hosted — no install. Add this as a custom connector in Claude, ChatGPT (developer mode) or any MCP client:

https://newspacemarket.com/mrd-connect/mcp

Local — for editors that speak stdio (Cursor, Windsurf, VS Code, Claude Desktop). Add to your MCP config:

json
{  "mcpServers": {    "mrd": {      "command": "npx",      "args": ["-y", "@newspacemarket/mrd-mcp"]    }  }}

The package is a thin bridge: it speaks stdio to your editor and forwards to the hosted endpoint, so new tools appear without upgrading anything.

Eight tools arrive in the assistant:

toolwhat it does
match_componentsreverse search — requirements in, candidates grouped A / B / C with per-parameter verdicts; optional price ceiling and export-control filter
get_component_specsthe full typed layer of one category, for your own comparison or trade study
build_budgetmass and steady-state power of a stack, with bus-voltage checks and an honest list of what could not be counted
get_datasheet_digestfacts machine-read from one product's datasheet, by section (electrical, interfaces, mechanical, thermal, qualification, commercial…), each with page and verbatim quote
search_datasheet_factsone fact key across a whole category — connector, protocol, TID, flight heritage, lead time, price — one row per product with page and quote
ask_datasheetask a datasheet a question and get the closest excerpts with page numbers, or scan every datasheet of a category for an exact term
design_stacka whole spacecraft bus from a plain-language task; pass fixed product ids to lock parts you have already chosen
find_data_gapswhat is missing for a product, company or category and where to get it: already in the digest, in the datasheet text, or only from the vendor

Where it is listed

  • Official MCP Registry — io.github.newspacemarket-com/mrd
  • Smithery — info-zrja/newspace-mrd, one-click install into Claude, Cursor, VS Code and other clients
  • Dataset mirror — Hugging Face

Use it from code

No authentication for reading. An optional free key raises rate limits five-fold — request one at /mrd-connect/.

bash
curl "https://newspacemarket.com/mrd-connect/match?category=star-trackers&accuracy_max=10&bus_v=12"

The same four datasheet-layer calls are plain REST as well: /mrd-connect/digest?product_id=…, /mrd-connect/facts?category=…&key=…, /mrd-connect/datasheet?product_id=…&q=…, /mrd-connect/gaps?category=….

OpenAPI 3.1: openapi.json, served live at https://newspacemarket.com/mrd-connect/openapi.json. Import it as a ChatGPT Custom GPT Action, a LangChain / CrewAI / n8n toolkit, or convert it to Gemini function declarations.

Download the data

data/ holds a dated snapshot, regenerated as the layer grows:

filecontents
mrd-full.jsoneverything — products, values, property dictionary
mrd-values.csvone row per value: product, property, number, unit, bound, condition, source, confidence
mrd-products.csvone row per product: name, manufacturer, category, URL
mrd-properties.csvthe property dictionary — what each key means and its canonical unit

Live data is always newer than the snapshot; use get_component_specs or /mrd-connect/specs?category=… when freshness matters.

How the data is built

Values are extracted from vendor datasheets (text and, for scanned documents, page images), from vendor product pages, and from product descriptions — the source of every record says which. Units are converted by code, not by a model: kg to g, dBm to W, psi to bar, tf to mN. A validation pass rejects a figure whose unit does not match the property, and an outlier scan flags values that sit far from the category median for human review.

Coverage is not exhaustive. Where a vendor publishes nothing machine-readable, the product simply is not here — and where a required field is missing, the API says so by name instead of guessing.

Licence

  • Code (the MCP bridge and examples): MIT — see LICENSE.
  • Data in data/: CC BY 4.0, attribution "NewSpace Market" — see data/LICENSE-DATA.

Specifications are extracted from publicly published vendor materials and are provided as-is. Verify against the vendor datasheet before committing to a design.

Links

Are you a vendor whose products are listed? Completing the missing fields takes minutes and is free — write to us from your company domain and we will send a direct link.

来源:README.md,提交 e2bc70b

工具

14

explain_properties

What every property of a product category means: definition, how it is measured and what goes into the measurement condition, how to use it when selecting, what it is commonly confused with, typical range and related properties. Call this before interpreting specs of an unfamiliar category.

输入参数subtypecategory必填

match_components

Reverse search over a typed spec layer: pass a category and engineering requirements, get every catalogued item grouped as A (meets all stated requirements), B (fails a named requirement, both figures quoted) or C (could fit, data missing). Verdicts cite the datasheet page. Omit parameters you do not care about. Categories: reaction-wheels, star-trackers, obc-computers, batteries, radios, solar-panels, propulsion, sun-sensors, magnetometers, magnetorquers, imus-gyros, gnss-receivers, eps, deployers, thermal, antennas, sdr-modems, cameras, structures, platforms, ground-stations, sadm, tanks-valves, separation-systems, pointing-mechanisms, test-facilities (environmental test labs — items are companies: test, shaker_min_kn, tvac_min_m, specimen_min_kg, emc_min_m). The answer is compact: group A in full (up to 25), groups B and C — the first 10 with the reason, plus counts and the fields most often unpublished in C; page with group and offset.

输入参数bandtestbus_vgrouplimitmediaorbitexportform_uoffsetacq_maxarw_max另有 78 个

find_alternatives

Alternatives / second sources / replacements for ONE product: requirements are derived from its own published values (key fields of its class — higher-is-better at least 80 %, lower-is-better at most 120 %, same interface, band or type) and its class is searched without it. Returns the derived requirements, fields not compared, and groups A (meets all), B (fails one), C (data missing). Use when a part has a long lead time, is discontinued or export-restricted. Equal figures do not make a drop-in replacement — interfaces must be confirmed with the manufacturer.

输入参数countryproductproduct_idtolerance_pct

selection_guide

How to choose a product of one category, from the catalogue data only: the decisive parameters with real ranges (middle half and full range), which side is better, how many products publish each, what to ask the vendor, and ready match queries. category = "cubesat" gives the list of subsystems a CubeSat is built from with catalogue counts.

输入参数category必填

find_suppliers

Companies whose PRODUCTS meet engineering requirements in one category — "who makes platforms for GEO", "Chinese companies making star trackers better than 10 arcsec", "suppliers of flight-proven reaction wheels". Pass category, optional country and any requirement parameter of match_components (same names and units). Returns companies with their matching products, and how many more companies may fit because the decisive value is not published.

输入参数bus_vorbitexportcountrydata_ifcategory必填mass_maxprop_typeaccuracy_maxmomentum_minflight_provenpaymass_min_g另有 3 个

build_budget

Engineering budget for a component stack: pass product ids (from the other tools) and your bus voltage, get total mass, total steady-state power, per-item data interfaces, voltage-compatibility verdicts and an honest list of what could NOT be counted because the vendor does not publish it. Use it to assemble and sanity-check a subsystem, e.g. 3-4 reaction wheels for an ADCS.

输入参数bus_vitems必填

get_component_specs

The typed layer of a category, 10 products per call (page with offset, narrow with ids and keys) — every item with every value (number, canonical unit, bound, measurement condition, datasheet source page, confidence L0/L1) plus the property dictionary grouped by interface port (perf/power/data/mech/thermal/env/supply) — for arbitrary comparison, budgeting or trade studies on your side.

输入参数idskeyslimitoffsetcategoryconfidence

get_datasheet_digest

Facts machine-read from a product datasheet, by section (identity, performance, electrical, interfaces, mechanical, thermal, environmental, qualification, lifetime, options, compliance, commercial). Each fact carries unit, condition, the page and a verbatim quote it was verified against (confidence L0) plus a link into the PDF. Use it for what the typed layer does not hold: connectors, protocols, survival temperature, vibration, radiation, heritage, options, export control, lead time. Pass product_id from match_components or a product name.

输入参数qproduct_id

search_datasheet_facts

One fact key across a whole category: which products state it in their datasheets and what value — e.g. connector, data_interface, protocol, survival_temperature, tid_tolerance, flight_heritage, trl, export_control, lead_time, price, design_life. One row per product with page and quote. Reports coverage (how many products of the category have a digested datasheet).

输入参数key必填category必填contains

ask_datasheet

Ask a product datasheet a question and get the closest excerpts with page numbers (pre-indexed text), or scan ALL datasheets of a category for an exact term (term + category) and get the exact list of products mentioning it with an excerpt each.

输入参数qtermcategoryproduct_id

design_stack

Assemble a whole spacecraft bus from a plain-language task: the service plans the subsystems, matches every item deterministically against the stated requirements, computes the mass/power budget and writes an evidence-bound rationale. Pass fixed = product ids you have already chosen (from match_components or get_datasheet_digest): they stay in their subsystems, are checked against the requirements (selected.fixed, fixed_verdict, fixed_fails) and the rest is assembled around them. LLM-backed and rate-limited (6/min, 30/h anonymous).

输入参数q必填fixed

find_data_gaps

What is missing for a product, a company or a category and where to get it: in_digest (already read from the datasheet — value, page), in_datasheet_text (the datasheet mentions it — page, excerpt), not_in_datasheet (ask the vendor), no_datasheet (ask the vendor for a datasheet). For vendors: what to publish; for buyers: what to ask.

输入参数statuscategorycompany_idproduct_id

ask_console

Plain-language engineering question over the whole platform, answered by platform rules without any external model: selection by requirements, values of a property across a class, companies by capability and country, launch windows, term definitions. Returns a short answer plus data blocks (candidates A/B/C, values, companies, launches). Use when the question does not map cleanly onto one structured tool.

输入参数q必填

check_fit

Check whether a set of products docks together, by the interface template of each product type: deployer and CubeSat structure form factor and satellite mass, separation system and spacecraft mass, solar panel size and EPS solar input (Voc/Vmp, power), battery and EPS, EPS output rails or the unregulated bus against each unit supply range, EPS charge current against the battery limit, satellite envelope against the deployer envelope for its form factor, on-board computer against unit data interfaces, board stack format, free volume, radio against antenna and ground station (frequency, band, connector, protocol and modulation), antenna against ground station polarization, GNSS receiver signals against the GNSS antenna band, magnetorquer coil current against the driver channel, PPU output power against the electric thruster, array drive against panel power, propulsion propellant against tank media, steady power budget. Verdicts fits / conflict / check / unknown, plus ask_vendor: exactly which values are missing to decide.

输入参数items必填

版本历史

1
  1. v1.2.0最新Sep 16, 2026

评分

已满足适用检查项的 100%

该评分衡量此收录条目的完整程度,依据注册中心条目、端点与公开来源计算得出。它不是对服务器本身的评价,也不表示服务器安全或当前可用:可连通性是带日期的检查结果,评分也不包含安装量或用户评价。

  • 近 30 天内可连通必需25 分
  • 已声明认证方式必需10 分
  • 已列出工具必需15 分
  • 描述必需15 分
  • 图标10 分
  • 文档15 分
  • 许可证10 分