SnapEDA Parts Library

SnapEDA Parts Library - maintained by SnapEDA under its newer name SnapMagic Search - is the Electronic Components slice's design-content layer: roughly 10 million parts from 2,500+ manufacturers, each anchored on an MPN/manufacturer pair and dressed with whatever the maker supports beyond the bare number. That means schematic symbols, PCB footprints and 3D models exportable to about twenty EDA tools, plus the surrounding context engineers actually shop with - catalog description, indicative price, datasheet linkage and manufacturability checks. The platform claims a stated 2 million engineer users, which makes it less a website than the default reference layer between a datasheet and a placed order.

Two structural facts make it a dataset instead of a parts bin. First, the keying is clean: one record per MPN/manufacturer pair, no distributor aliases muddying joins, so the same 0740991012 + Molex pair resolves identically whether it arrives from a BOM tool, a price feed or this library. Second, the library moves. Its New Parts Added feed shows entries minutes old - an Infineon automotive MOSFET at one hour thirteen, a Molex connector housing at one hour thirty-seven during verification - which turns library entry into a timestamped event rather than a static attribute.

One part from the New Parts Added feed, read straight off its result card:

mpn        : IAUC120N04S6L012ATMA1   manufacturer: Infineon Technologies
time_added : 1 hour, 13 minutes      cad_models  : true
description: OptiMOS power MOSFET, automotive N-channel, PG-TSDSON-8
price_usd  : (unlisted on card)      cad_formats : Altium, KiCad, Eagle +more

mpn        : 0740991012              manufacturer: Molex
time_added : 1 hour, 37 minutes      cad_models  : true
description: MICRA CLIK-MATE 1.25mm pitch, 12-pos receptacle housing
price_usd  : 1.227                   cad_formats : Altium, KiCad, PADS +more

Note what the second row proves: pricing appears where the maker or channel supplies it and stays absent where it does not, and both rows still resolve to complete, joinable records. The cad_formats column compresses a per-tool inventory into one field - roughly twenty export targets from Altium and KiCad through Allegro, Eagle, DipTrace, PADS and Fusion - because the format decision belongs at download time, not at indexing time.

Eight fields carry each part record, keyed on the MPN/manufacturer pair. Every definition traces to the source's own result-card layout and product copy rather than guesswork, though it bears saying plainly: the deep part pages sit behind bot protections, so these definitions are inferred from observable surfaces, not read off page internals. Six of the eight are stable across essentially the whole catalog; the datasheet link and the indicative price populate where the source lists them, and the three richer surfaces described below ride along on a part-dependent subset, so their observed values fold under additional fields on request.

The dictionary is deliberately shallow-and-clean rather than deep-and-messy. Where sibling sources in this slice publish nested per-seller offer ladders, this library publishes one part, one identity, one set of design assets - which is precisely why it joins without remapping against distributor catalogs, lifecycle feeds and supply-chain risk scores keyed on the same MPNs.

• Geography - global by construction. The library mirrors worldwide manufacturer catalogs with no regional scoping: US semis, European power and sensor houses, Asian connector and passive franchises all land in the same grid, which makes it a natural neutral frame for origin-diversification work alongside the slice's China-origin risk score.
• Temporal - continuously refreshed. The New Parts Added feed showed minute-level entries during verification, and the archive reaches back through years of accumulated catalog support. No other source in this industry slice pairs ten-million-part scale with hour-fresh arrivals.
• Granularity - one record per MPN/manufacturer pair with its CAD assets attached, aggregating naturally up to manufacturer-level views across all 2,500+ makers. Manufacturer rollups answer a question almost nothing else answers directly: which component makers invest in design-enablement content, and how recently did they last bother.

The honest caveat: depth is not uniform. Asset coverage - symbol, footprint, 3D model - varies part to part, and the availability flag is the field that keeps that variation queryable instead of surprising.

API, files, or your warehouse. Daily, weekly, or hourly.

You pick the channel and the cadence; the field dictionary above travels unchanged through all three. Part-page HTML never reaches your pipeline - rows arrive flattened to one observation per MPN/manufacturer pair with identities already normalized, so your BOM keys line up against ours without a fuzzy-matching pass. Cadence changes are a settings conversation, not a re-integration project, and a sample cut to your MPN list comes first either way.

• Data scientists and ML engineers mine spec, datasheet and download metadata across the ~10M-part corpus for component-recommendation models, BOM intelligence features and training sets keyed on clean MPN identity - relevance 2 in their pack, and the largest single text-plus-metadata corpus in the slice.
• Competitive-intel and product teams compare symbol, footprint and 3D-model coverage across 2,500+ manufacturers to gauge which rivals are winning the design-in battle before purchase orders say so - also relevance 2.
• Developers and builders integrate ECAD-adjacent features - part search, model availability checks, format routing - on top of a stable per-part schema instead of maintaining scrapers per manufacturer.
• Sales growth teams treat CAD support breadth by manufacturer as a proxy for design-in activity, prioritizing accounts whose makers are visibly investing in enablement content.
• Investors and quants watch the minute-level new-parts feed as an early, public design-activity signal - a leading indicator that shows up in library timestamps well before it lands in quarterly revenue.

Data scientists hold this at relevance 2 in their electronic-components pack - the ~10M-row corpus is the biggest labeled-by-maker body in a slice otherwise dominated by transactional catalogs. Competitive-intel teams match at relevance 2: CAD coverage is a design-in signal competitors rarely think to monitor systematically. Developers and builders also score 2, building part-search and ECAD-integration features on a fixed vocabulary. Sales growth teams and investors/quants hold relevance 1 but use it in ways nobody else can - reading enablement investment and design activity out of a parts database. The constant across all five: the MPN key. Everything this library offers compounds when joined against the slice's price, stock and lifecycle feeds on the same part numbers.

Four things, all knowable upfront. First, provenance honesty: field definitions here are inferred from the source's observable surfaces - result cards, homepage copy, published API descriptions - because deep part pages sit behind bot protections; the definitions are sound, but a sample verifies them against your actual parts before you build. Second, sparsity is real: prices and datasheet links populate where the source lists them, not everywhere, and CAD asset coverage varies part to part - the availability flag exists precisely so you can scope requests to fully-covered subsets. Third, prices are indicative display figures from result cards, not negotiated distributor pricing; teams modeling realized cost should join against a distributor-offer feed rather than treating this column as revenue truth. Fourth, the underlying design files carry their own terms at the source level, while Datadory delivers derived, normalized rows under our own terms - the distinction matters for redistribution planning and is spelled out in the sample paperwork before anything ships.