Analytics cookies only (Google Analytics, PostHog). No ad networks, no cross-site tracking. See our privacy policy.
upstream: github.com/inventree/InvenTree (opens in new tab) · github.com/Part-DB/Part-DB-server (opens in new tab)
Pick InvenTree
If you need purchase orders, build orders, a REST API and plugins under an MIT licence, and are fine with a four-service install.
Pick Part-DB
If you manage electronic components, want KiCad and supplier lookups, and prefer one container on SQLite with the lighter footprint we measured (~180 MB against ~697 MB).
InvenTree and Part-DB are the two parts-inventory systems in our catalog, and both handle electronic parts well. Part-DB is a Symfony (PHP) app aimed squarely at electronic components, with footprints, KiCad integration and supplier lookups, and it runs happily on SQLite in one container. InvenTree is a Python and Django system with a React front end, a REST API, plugins and a background worker, which its docs describe as designed for "SME or hobbyist applications". Pick based on whether you want a focused components database or a broader stock and supply system.
Part-DB's documentation says it is "an Open-Source inventory management system for your electronic components." Each part "can be assigned to a category, footprint, manufacturer, and multiple store locations and price information." It generates barcodes and labels, has a webcam barcode scanner, and offers users and groups with "detailed (fine-grained) permissions". Projects can be given a bill of materials from your parts. It can fetch part data from providers such as Octopart, Digikey, Farnell, Mouser and TME.
InvenTree's docs call it "an open-source inventory management system which provides intuitive parts management and stock control." Parts live in categories; stock items live in nested locations and can be serialized. "Intelligent BOM (Bill of Material) management" tracks sub-parts, suppliers can be linked to any part, and build orders track progress while consuming stock.
Part-DB advertises that you can "use Part-DB as the central datasource for your KiCad and see available parts from Part-DB directly inside KiCad." Its docs also cover two-factor authentication (Google Authenticator and WebAuthn keys) and optional SAML single sign-on, for example through Keycloak. InvenTree has its own TOTP two-factor login and OpenID or OAuth single sign-on through django-allauth, so login security is not the deciding factor.
InvenTree's feature set leans toward manufacturing and procurement: its purchase orders "allow to track which parts are bought from suppliers and manufacturers", and its build orders and BOM management are core features. For a hobbyist with a shelf of resistors, Part-DB's component-first model is the closer fit; for tracking assemblies you build and parts you order, InvenTree covers more.
InvenTree runs on a RESTful API with a native Python binding library, and has a plugin system with built-in plugins and plugins from PyPI or a local directory. Its docs note that third-party plugins are not supported or guaranteed.
Part-DB provides an API for accessing it from other applications and scripts, and its info providers pull part data into your records.
Part-DB supports SQLite, MySQL/MariaDB and PostgreSQL as database backends. Its Docker image defaults to SQLite, and its Docker guide documents a MySQL variant; ours uses SQLite, with a single container and three bind-mounted directories. InvenTree's Docker setup runs a server, a background worker, PostgreSQL and a Caddy proxy that serves static and media files.
Our page lists 256 MB for Part-DB and 1 GB for InvenTree as minimums; neither project publishes a RAM figure, so these are our estimates. In our install check (GCP e2-standard-2, Ubuntu 26.04), idle at first boot with the default install snippet and no parts added, Part-DB used ~180 MB of RAM and ~1.0 GB of disk, and InvenTree used ~697 MB of RAM and ~2.7 GB of disk (both November 2026). Those are idle figures; importing a large catalogue, running the worker or generating labels will use more on both.
Part-DB is 2 out of 5: one container, an APP_SECRET, and a first-login
password to change. InvenTree is 3 out of 5: four services, a Caddyfile, an
admin password and site URL settings that must match how you reach it.
Choose Part-DB if you manage electronic components, want KiCad and supplier lookups, an AGPL-3.0 app that runs on SQLite, and the lighter footprint.
Choose InvenTree if you need purchase orders, build orders, a REST API and plugins, a permissive MIT licence, and are fine with a heavier four-service install.
Both installs are verified on Ubuntu 26.04; the pinned snippets are on the
InvenTree and Part-DB pages. Part-DB
fits the smallest plan; give InvenTree 1 GB or more. Back up the Part-DB db,
uploads and public_media folders, or InvenTree's database and data
directory, and put either behind TLS.
Part-DB lists a lower minimum: 256 MB vs 1024 MB for InvenTree.
No. InvenTree is MIT, Part-DB is AGPL-3.0.
Part-DB for electronic components with KiCad integration, supplier lookups and a single-container install (~180 MB idle in our first-boot check). InvenTree for purchase orders, build orders, a REST API and plugins (~697 MB idle, with a server, worker, database and proxy).
Yes. Part-DB's docs say you can use it as the central datasource for KiCad and see available parts from Part-DB directly inside KiCad.
Yes. InvenTree's docs say the core software runs on a RESTful API that external applications can use, and a native Python binding library is also provided.
Part-DB supports SQLite, MySQL/MariaDB and PostgreSQL; its Docker image defaults to SQLite, which our install uses. Our InvenTree install runs PostgreSQL alongside a server, a worker and a Caddy proxy.
Paid link — we earn a commission if you shop through it.
A whiteboard-and-docs workspace on a plain Compose file vs. native apps on a commercial server.
A full identity platform vs. a passkey-only sign-in box.
Two modern IdPs, built for different jobs.