explaingit

replenhq/replen

Analysis updated 2026-05-18

7TypeScriptAudience · developerComplexity · 3/5LicenseSetup · easy

TLDR

Replen watches open source projects and your own repos locally, then surfaces a small number of relevant libraries, algorithms, or techniques your AI coding tool can use.

Mindmap

mindmap
  root((Replen))
    What it does
      Maps your repos
      Watches OSS frontier
      Scores relevant matches
      Suggests inside AI session
    Tech stack
      TypeScript
      Node.js
      MCP
    Use cases
      Library adoption
      Algorithm porting
      Dependency tracking
    Components
      Brainstem
      Watchtower
      Atlas

Code map

Detail Auto

An interactive map of this repo's files and how they connect — its source is parsed live in your browser. Click Visualize to build it.

filefunction / class

What do people build with it?

USE CASE 1

Get a short suggestion inside Claude Code or Codex when an open source project fits something you are building.

USE CASE 2

Triage today's matches for the current repo with a verdict and effort estimate per candidate.

USE CASE 3

Track releases, security advisories, and license changes for dependencies your code relies on.

USE CASE 4

Build a local knowledge graph of your projects, capabilities, and past adoption decisions.

What is it built with?

TypeScriptNode.jsMCP

How does it compare?

replenhq/replenadguardteam/ruleseditorarkorlab/haru
Stars777
LanguageTypeScriptTypeScriptTypeScript
Last pushed2026-07-01
MaintenanceActive
Setup difficultyeasyeasyhard
Complexity3/52/55/5
Audiencedeveloperdeveloperops devops

Figures from each repo's GitHub metadata at analysis time.

How do you get it running?

Difficulty · easy Time to first run · 5min

Requires signing in via Google, GitHub, or email link during the one-command install.

Apache 2.0 license: use freely for any purpose, including commercial use, with patent protection, as long as you keep the copyright notice.

In plain English

Replen is a local intelligence layer for AI coding tools like Claude Code, Codex, and Cursor. Instead of only helping inside the single repo you are working in, it builds a map of everything you build across your whole portfolio of projects, such as computer vision work, geospatial analysis, scraping, or workflow automation, and continuously watches open source projects to find things that could actually help your real code. Rather than showing a trending list or a keyword search, Replen scores open source projects against what your own repos actually do, and surfaces a small number of high signal suggestions on quiet days, sometimes none at all. A suggestion might mean adopting a library as is, porting a specific function or algorithm from a larger project, cherry-picking a technique, or building something from scratch in a clean room way to stay license compliant. This matching and judgment happens locally, inside your coding session, against your real code, and the README states that your code never leaves your machine. The system has three parts. Brainstem is the matching core that understands what each of your repos does through per-capability embeddings rather than keywords, and it learns from what you adopt or skip over time. Watchtower is a maintained set of about 1,250 sources tracking releases, security advisories, pricing, licenses, and end of life dates for the things your code depends on. Atlas is a knowledge graph of your projects, capabilities, and decisions, built from linked markdown files your AI agents can read directly. Setup is three steps: install with a single npx command that signs you in and asks which repos to include, run a one-time onboarding command inside Claude Code so Replen can read your repos and build a profile of what you actually build, and then continue working as normal, with Replen occasionally leaving a short note in your AI tool's reply when something relevant is found. Replen finds your repositories by looking for folders containing a git repo with a GitHub remote, checking several conventional locations before asking you directly if none are found. The full README is longer than what was shown.

Copy-paste prompts

Prompt 1
Show me how to install Replen with npx replen and onboard my first project with /replen-onboard.
Prompt 2
Explain the difference between Replen's Brainstem, Watchtower, and Atlas components.
Prompt 3
How does Replen decide whether to suggest adopting, porting, or clean-room building a piece of code?
Prompt 4
Help me set REPLEN_PROJECT_ROOTS so Replen finds repos in a non-conventional folder.

Frequently asked questions

What is replen?

Replen watches open source projects and your own repos locally, then surfaces a small number of relevant libraries, algorithms, or techniques your AI coding tool can use.

What language is replen written in?

Mainly TypeScript. The stack also includes TypeScript, Node.js, MCP.

What license does replen use?

Apache 2.0 license: use freely for any purpose, including commercial use, with patent protection, as long as you keep the copyright notice.

How hard is replen to set up?

Setup difficulty is rated easy, with roughly 5min to a first successful run.

Who is replen for?

Mainly developer.

Open on GitHub → Explain another repo

This repo across BitVibe Labs

Verify against the repo before relying on details.