muratg98

muratg98 / psychmem

Public

Psychology-grounded selective memory system for AI agents (Claude Code, OpenCode) with STM/LTM consolidation, decay curves, and learned importance weights

archived deprecated unmaintained
44
5
100% credibility
Found Feb 21, 2026 at 21 stars 2x -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
TypeScript
AI Summary

PsychMem adds human-like persistent memory to AI coding agents, automatically extracting and recalling important details like decisions, bugs, and preferences across sessions.

How It Works

1
šŸ” Discover PsychMem

You stumble upon PsychMem while searching for ways to make your AI coding buddy remember things better between chats.

2
šŸ“„ Add the Memory Helper

You simply copy the memory files into your AI coding tool's special folder, and it's ready to go.

3
šŸ§  Magic Starts Happening

As you chat with your AI about code, bugs, and ideas, it quietly captures and stores the key moments without you lifting a finger.

4
šŸ’¬ Keep Building Over Time

Days or weeks later, start a fresh conversation and watch your AI pull up past decisions, fixes, and preferences automatically.

5
ā­ Tweak What It Remembers

Use simple commands to pin favorites, forget extras, or boost important ones so it learns your style perfectly.

šŸš€ AI Becomes Your Memory Partner

Your coding sessions flow effortlessly now, with your AI acting like a smart teammate that never forgets the details.

Sign up to see the full architecture

4 more

Sign Up Free

Star Growth

See how this repo grew from 21 to 44 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 psychmem?

Psychmem is a TypeScript-based selective memory system for AI agents like Claude Code and OpenCode, drawing from psychology to give them human-like persistence across sessions. It extracts key details from conversationsā€”preferences, bugs, decisionsā€”scores them by learned importance weights, and stores them in STM/LTM with consolidation and decay curves modeled on Ebbinghaus forgetting. Developers get automatic memory injection at session start, so agents recall project context without constant re-explaining.

Why is it gaining traction?

Unlike basic context windows that reset everything, psychmem uses psychology-grounded rules for what sticks: multilingual patterns detect emphasis or errors, structural analysis spots corrections, and scoped retrieval separates user-wide from project-specific memories. The CLI lets you search, list, pin, or forget memories on demand, while hooks handle real-time extraction during tool use or compaction. At 19 stars, it's niche but hooks coders tired of amnesia in agent workflows.

Who should use this?

AI agent users building with Claude Code or OpenCode, especially on multi-session projects where forgetting API choices or bug fixes kills momentum. Backend devs debugging persistent issues, or teams enforcing coding constraints like "never use var," will notice fewer context switches. Skip if you're on single-shot agents or don't need cross-session memory.

Verdict

Worth a test install for Claude Code/OpenCode usersā€”detailed docs and research refs make setup straightforward despite 1.0% credibility and low maturity (19 stars, no tests visible). Early but functional; watch for stability as it evolves.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.