LittleYier

LittleYier / site-memory

Public

Persistent memory for browser automation AI Agents. Learns site knowledge automatically, updates after every visit, and gets better over time.

16
3
100% credibility
Found Apr 10, 2026 at 15 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
JavaScript
AI Summary

site-memory is a persistent memory system for AI browser automation agents that automatically learns and updates knowledge about websites to improve speed and efficiency on repeated tasks.

How It Works

1
🔍 Discover site-memory

You learn about a helpful memory tool that teaches your AI web assistant how websites work, making it faster on familiar sites.

2
🤖 Ask your AI to add it

Simply tell your AI helper to install the memory tool, and it handles everything automatically.

3
🧠 Memory is ready

Your AI now remembers website layouts and steps from past visits without any extra setup.

4
🌐 Give your first web task

Tell your AI to do something on a website like filling a form or searching, and it learns the site while completing the job.

5
🔄 Try the same site again

The next time you visit the same website, your AI uses its memory to finish much quicker and with fewer mistakes.

⚡ Speedy web automation

Your AI zips through repeated website tasks effortlessly, saving time and effort every time you use it.

Sign up to see the full architecture

4 more

Sign Up Free

Star Growth

See how this repo grew from 15 to 16 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is site-memory?

site-memory is a JavaScript skill for AI agents that adds persistent memory to browser automation tasks. It automatically learns and stores knowledge about websites—like selectors, navigation paths, and layouts—updating after every visit so agents get faster and more accurate over time. Developers get a drop-in layer stored at ~/.site-memory/ that cuts repeated work on sites like Booking or Amazon, solving the problem of agents re-learning site structures from scratch each run.

Why is it gaining traction?

It stands out with WebVoyager benchmarks showing 67-76% fewer commands and time on 15 sites after three rounds, using cheaper models like Claude Haiku while matching Sonnet accuracy. Easy install via npx skills add or AI prompts works with any browser tool, from Chrome CDP to Playwright MCP, and provides github persistent storage that feels like muscle memory for agents. For github persistent memory or browser-use setups, it's a low-effort win over stateless alternatives.

Who should use this?

Agent builders automating repetitive web tasks, like filling insurance forms on example-insurance.com or searching apartments on realestate-site.com. Teams using Claude Code, Gemini CLI, or Codex for e-commerce scraping, admin dashboards, or benchmarks like WebVoyager. Ideal for devs needing github persistent identifier for sites that change slowly.

Verdict

Try it if you're in browser agent workflows—benchmarks deliver real speedups, docs are solid in English/Japanese/Chinese, MIT license. At 12 stars and 1.0% credibility, it's early but promising; test on your sites before production.

(178 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.