Blog / Comparisons / FIG. 149
Jev Directories, Compared Honestly
Every major Jev directory compared, including ours: what each catalogs, how big it says it is, and which question each one answers best.
Pick any Jev directory and you'll find hundreds of links. Jev, the decision model from TypeSafe AI, launched in September 2026, and within two weeks the ecosystem had more lists of builds than some platforms get in a year. This site is one of them, so here's the obvious objection up front: a directory comparing directories is grading its own homework. Fair. We'll state what the others do better and let you check.
The comparison
Sizes below are what each directory states about itself on its own page at the time of writing; they change daily, so treat them as snapshots, not scores.
| Directory | Format | What it catalogs | Stated size | Best for |
|---|---|---|---|---|
| hellogumbo/awesome-jev | GitHub list plus search site | Projects, SDKs, integrations, agent tooling, articles | 1,094 entries as of September 24, per its README | Breadth, and SDKs across many languages |
| AbdelStark/awesome-typesafe-jev | GitHub list | Source-backed field guide: SDKs, demos, agent tools, independent evaluations | Not stated as a count | A curated tour with evaluations called out |
| fatwang2/awesome-jev | GitHub list | Source-backed project directory with a GitHub review workflow | Not stated as a count | Seeing how entries get reviewed |
| jevdirectory.org | Website | Tools, repos, guides, cookbooks, patterns, a glossary | 638 entries from 102 source sites, per its homepage | Learning resources and tooling |
| shipwithjev.com (this site) | Website | Builds with author-reported numbers and sources | Several hundred builds | What a build did and what it cost |
Several GitHub lists share the awesome-jev name; our awesome-jev explainer covers the genre and why awesome lists index existence rather than quality.
Where the others beat us
Breadth, clearly. If you want an SDK in an unusual language, an MCP server in Go, or an article thread from day three, the big awesome lists and jevdirectory.org will have it and we may not. They catalog tools and resources; we catalog builds. We also have no independent evaluation section of the kind some field guides maintain. When you're choosing a library, start with them.
Where we're different
Receipts. Every entry here aims to carry the builder's own numbers (items processed, cost, time) with the source linked and "as reported" attached, because none of these numbers are audited and no official Jev benchmarks exist. How we collect, hedge, and correct those numbers is spelled out in the receipts methodology. A typical entry reads like the 500 emails classified for about 3.5 cents build, as reported by its author, not like a link with a star count.
That's a narrower promise than "everything," and deliberately so. A list answers "what exists?" A receipt answers "what did it do, and what did it cost?" You probably need both.
How to use them together
Discover in the broad lists, verify in the source repo, and check the receipts before you copy an architecture. When something you find works, submit it here with its numbers. And if you're still orienting on what Jev is, no directory replaces the official docs at docs.typesafe.ai.
Frequently asked questions
What is the biggest Jev directory?
By stated count at the time of writing, hellogumbo's awesome-jev, at 1,094 entries per its README. Counts move daily and measure different things.
Which Jev directory is official?
None of them. TypeSafe's official source is docs.typesafe.ai; every list here, including this site, is independent.
Why does shipwithjev list fewer entries?
We catalog builds with reported numbers rather than every tool and article, per our methodology. Fewer entries, more context each.
Are the numbers in these directories verified?
Mostly not independently: they're builder-reported. Read them as "as reported" and test on your own data before relying on them.
Numbers throughout are as reported by the build authors, not verified by shipwithjev. Code-shaped examples are pseudocode; the official docs live at docs.typesafe.ai.