IncomeStreamSurfer

IncomeStreamSurfer / paperclip-surfers

Public

Self-improving AI agent companies — Paperclip V2 fork with persistent memory, MCPs, skills management, and KPI tracking. Built by HarborSEO (harborseo.ai)

16
9
100% credibility
Found Mar 31, 2026 at 16 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
TypeScript
AI Summary

Paperclip Surfers is a user-friendly dashboard for creating and managing self-improving teams of AI agents with persistent memory, skills, and performance tracking.

How It Works

1
🔍 Discover Paperclip Surfers

You find this tool on GitHub for building and managing teams of smart AI helpers that work together like a company.

2
💻 Set it up on your computer

Download the files and prepare everything with simple steps—no coding needed.

3
🚀 Launch your AI company dashboard

Click to start and see your personal control center come alive instantly.

4
🧙‍♀️ Follow the friendly setup guide

Answer easy questions to connect AI thinkers and set basic preferences.

5
👥 Build your AI dream team

Add roles like CEO, developers, and specialists with special skills.

6
📋 Give tasks and watch magic happen

Assign projects and see your team collaborate, learn, and improve over time.

🎉 Your AI company runs smoothly

Relax as your self-improving team handles work, remembers lessons, and grows smarter.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 16 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 paperclip-surfers?

Paperclip-surfers is a TypeScript-based control plane for running self-improving AI agent companies, forking the original Paperclip project with additions like persistent conversations, agent memory across sessions, MCP integration from Claude Code, per-agent skills selection, and KPI tracking for self-improvement loops. Built by HarborSEO (harborseo.ai), it lets you manage agent teams that resume Claude Code sessions, accumulate global or project-specific knowledge, assign tools like Slack or AWS, and monitor performance metrics like completion rates and costs. Developers spin it up via pnpm install/dev or Docker, accessing a dashboard at localhost:3100 for task assignment and analytics.

Why is it gaining traction?

It stands out by enabling persistent agent state—no more resetting conversations per task—plus one-click MCP sync from ~/.claude.json and toggleable skills from ~/.claude/skills, making multi-agent setups feel seamless for self-improving agentic AI systems. The self-improvement loop with post-run KPIs, CEO observations, and A/B experiments turns raw agent runs into trackable, evolving companies, echoing ideas from self-improving agents papers. CLI tools like `paperclipai onboard`, `doctor`, and `run` simplify setup and diagnostics.

Who should use this?

AI engineers building agentic teams for devops, SEO automation, or RAG pipelines will appreciate the memory injection and tool management. Teams at HarborSEO or experimenting with self-improving agent cloudbots can fork and extend it for custom KPIs. Solo devs prototyping embodied foundation models or Stanford-style self-improving systems get quick persistent state without rebuilding from scratch.

Verdict

Try it for agent experiments if you're okay with early maturity—16 stars and 1.0% credibility score mean light testing and docs, but solid Docker quickstarts and MIT license make forking viable. Pair with production monitoring before scaling agent companies.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.