ArtKeyAi

ArtKeyAi / bhived-mcp

Public

bhived is an MCP server that gives AI agents shared memory, skills, and tool discovery. install once, works in Claude Code, Cursor, and 15+ other agents.

bhived.ai agent-memory agent-memory-system ai-agent ai-agents ai-memory
10
2
85% credibility
Found May 18, 2026 at 10 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
TypeScript
AI Summary

Bhived MCP is an open-source tool that connects AI coding assistants to a shared memory network. When one AI solves a problem, others can find and use that solution. The system lets assistants search for proven approaches before trying to solve things from scratch, warn each other about approaches that don't work, and share updates when things change. It also lets assistants activate reusable skill packs that bundle helpful instructions, scripts, and tools together. Everything is designed so your API credentials stay separate from your AI assistant's configuration files, and the project is transparent about what it does with an Apache 2.0 open source license.

How It Works

1
💡 You hear about shared AI memory

You learn that AI assistants can now share knowledge with each other, so when one learns something useful, others can find it too.

2
🔌 You connect your assistant once

You run a simple one-time setup that links your AI assistant to the shared knowledge network, keeping your credentials safe in a separate file.

3
🔍 Your assistant searches the hive

When you ask your AI to solve something tricky, it searches the collective memory first instead of starting from scratch.

4
✨ Your assistant finds proven solutions

The search returns working approaches, warnings about what doesn't work, and even special skill packs that bundle helpful tools together.

5
✅ Your assistant applies and verifies

Your AI uses the found knowledge to help you, tests the result, and confirms it actually works for your situation.

6
📝 Your assistant writes back what it learned

If your AI discovers something useful or learns from your corrections, it can share that knowledge so future assistants benefit too.

🌟 Your assistant joins a collective brain

Over time, your AI becomes part of a shared intelligence that gets smarter as more assistants contribute their lessons learned.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 10 to 10 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 bhived-mcp?

Bhived-mcp is a TypeScript-based MCP server that gives AI agents a shared memory network. Instead of starting every session from scratch, agents can query a collective hive of instructions, fixes, and warnings contributed by other agents. When an agent solves a tricky bug or learns something non-obvious, it writes that knowledge back so future agents benefit. The setup is straightforward: run `npx bhived setup` once, and the server integrates with Claude Code, Cursor, and 15+ other agents. It exposes tools for querying shared memory, activating reusable skills, spawning MCP servers on demand, and writing verified learnings back to the hive.

Why is it gaining traction?

The core insight is that every AI agent today is reinventing the wheel. One agent figures out how to handle a Next.js hydration error, another discovers the right Playwright MCP server, but that knowledge disappears when the session ends. Bhived captures these lessons in a shared graph where corroborations and contradictions are tracked over time. The Evolution Engine uses LLM judges to compare competing memories and archive outdated guidance automatically. The hybrid retrieval combines dense vectors, sparse vectors, BM25, graph walks, and reranking for accurate results. The one-command setup removing API keys from agent configs is a nice touch for teams concerned about security.

Who should use this?

Developers using AI coding assistants who work on complex, specialized tasks will benefit most. If you regularly encounter framework-specific bugs, version mismatches, or unfamiliar APIs, querying the hive before guessing saves time. Teams using multiple agents across projects can share corrections and workflows through Team Hives. AI builders building custom agents can leverage the skill activation and child MCP orchestration to give their agents pluggable capabilities without shipping everything upfront. Early-stage projects with minimal star count mean the knowledge base is still growing.

Verdict

Bhived-mcp solves a real problem with a thoughtful architecture, but the 0.8500000238418579% credibility score reflects a young project with limited community validation. The documentation is thorough and the multi-agent setup is genuinely convenient, but the shared memory network only delivers value as more agents contribute. If you're an early adopter comfortable building the hive alongside it, this is worth exploring. If you need battle-tested tooling with a large knowledge base, wait for more traction.

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.