Memory & Context: what the data shows
All 29 projects in this category, what the distribution actually looks like, and which ones are still alive.
Memory is the single feature that separates a useful agent from a stateless chatbot, and it is also the hardest to get right. Every project here solves it differently, and the approaches are not compatible. Pick one early — migrating memory between systems is painful.
The numbers
| Measure | Value |
|---|---|
| Projects in category | 29 |
| Combined stars | 20,553 |
| Median stars | 161 |
| Highest | 12,053 |
| Pushed since Aug 2026 | 26 (90%) |
| Untouched since Jan 2026 | 1 |
Language mix: JavaScript (12), TypeScript (11), Python (6).
Licensing: MIT (23), NOASSERTION (2), Apache-2.0 (1).
The median project in this category has 161 stars while the largest has 12,053. That gap is the whole story: Most of what exists is someone's weekend experiment, published and forgotten. The handful at the top are not typical — they are outliers that happened to be found.
This matters when you choose. Ranking by stars is the only signal available across the whole ecosystem, but it is measuring attention, not quality, and it collapses to zero for anything new. A project with four stars and a push yesterday may serve you far better than one with four thousand and a last commit in March. That is why every entry below also carries a freshness marker.
Ranked shortlist
Sorted by stars, with a freshness marker: active means a push since August 2026, dormant means no recent activity.
| # | Project | Stars | Lang | License |
|---|---|---|---|---|
| 1 | MemOS active | 12,053 | TypeScript | Apache-2.0 |
| 2 | memsearch active | 2,782 | Python | MIT |
| 3 | dsh-mnemon active | 671 | TypeScript | MIT |
| 4 | graph-memory active | 662 | TypeScript | MIT |
| 5 | memtrace-public active | 505 | Python | NOASSERTION |
| 6 | dsh-memory active | 459 | Python | MIT |
| 7 | scope-recall-hermes active | 456 | Python | MIT |
| 8 | dsh-memory-evolve active | 428 | JavaScript | MIT |
Install the top pick
MemOS is the most adopted project here at 12,053 stars, written in TypeScript under Apache-2.0.
git clone https://github.com/MemTensor/MemOS.git
cd MemOS
npm install
npm run buildCheck its README for the exact registration step — the plugin entry point differs between projects and a wrong name fails silently in most setups.
What to skip, and why
1 of the 29 projects here have not been pushed to since January 2026. Star count did not save them — several are among the higher-starred entries in the category. These are the ones to check before you build a workflow around them:
yg33717/DSH_tensorflow— 105 stars, last push 2017-11-17
A dormant project is not automatically broken. It is a project where you will fix your own bugs.
Common questions
How many DSH plugins are in this category?
29, out of 714 catalogued across the whole directory. The median project here has 161 stars.
Which one should I install first?
MemOS, on the strength of community adoption rather than our own testing — we index the ecosystem, we do not benchmark it. If it does not fit your workflow, work down the shortlist above; the gap between first and fifth place is usually narrower than star counts imply.
Are these safe to install?
They run with the runtime's permissions, which usually means access to your files and API keys. Read the source of anything you install, especially in the categories that touch the shell or the browser. No directory can make that decision for you.
How often is this page updated?
The dataset behind it refreshes daily from the GitHub API. This article is regenerated whenever the site is rebuilt, so the numbers above are current as of 2026-10-24.
Browse the full category
Every project, filterable by language and license, lives on the Memory & Context category page. For the wider picture across all 714 projects, start at the full plugin index.