Orb marketplace and Builder licensing

Discover, share, and run MCP Orbs

Upload each Orb to get 90 free trial days — they stack cumulatively with no upper limit. Download the Builder for your OS, or get a license to publish.

Browse the catalog Sign in & upload

Store overview Live

30 days first download trial

90 days trial after upload

$99/yr Builder license

Search Orbs by tag, keyword, or retrieval method

hybridKnowledge Orb

vectorResearch Pack

bm25Local Docs

Start with Builder

Download the desktop Builder, then publish or activate from here.

⚡ Register & get trial ⬇︎ Builder Wizard for macOS

⬇︎ Builder Wizard for Windows ⚠ Unsigned — may trigger Windows SmartScreen warning (why?)

⬇︎ Builder Wizard for Linux

Advanced downloads (CLI, runtime, checksums)

Platform paths

Three entry points instead of one long checklist.

Discover

🔍 Browse Orbs

Search published Orbs by keyword, tag, or built-in search method. Download the original asset bundle or a ready-to-run Orb for your OS.

Browse the catalog

Publish

⬆︎ Upload an Orb

Sign in with your email, upload an Orb asset bundle, and we validate it automatically. Every successful upload can mint a free 90-day trial code.

License

Builder licensing

Purchase, activate, manage subscriptions, migrate devices, or recover your current Builder license from one dedicated center.

Open license center

Creator workflow

From local Builder to published Orb, with licensing handled alongside the store.

Build locallyPackage your knowledge assets with MCPOrbBuilder.
Upload and validateThe store checks structure and smoke-tests the asset bundle.
Share and licenseUsers download Orbs while creators keep Builder activation self-service.

What is MCPOrb?

The PDF for AI-native knowledge delivery.

MCPOrb packages product docs, playbooks, and expert knowledge into portable Orbs that run locally inside AI tools — no cloud, no data upload. The Store is where you browse, publish, and license them.

Learn more at mcporb.com → View on GitHub

PortableOne file: knowledge, index, runtime
LocalRuns on-device, no data upload
InspectableBuilt-in Web UI shows every source
MCP-nativeWorks in Claude, Cursor, VS Code

Real developer problems MCPOrb solves

How to load local documentation into Cursor securely without data leakage?
What is the best way to run MCP servers locally on macOS without cloud dependency?
Is there a portable MCP server that works with Claude Desktop offline?
How to package a company knowledge base as an MCP server for VS Code?

Technical Specifications

💻

Host Compatibility

Apple Silicon (M1/M2/M3 MacBook Air/Pro), Linux (Ubuntu 20.04+, Debian), Windows x64

🤖

AI Client Compatibility

Claude Desktop, Cursor, VS Code, and any MCP-compatible client

⌨️

CLI Commands

mcporb build — mcporb inspect — mcporb test-query — mcporb run

🔍

Search Methods

Hybrid / Vector / BM25 — built-in, no external service needed

🔒

Data Security

100% local. No data upload. Zero network exposure. Open-source code at GitHub for public audit.

Frequently Asked Questions — AI & Search Optimized

How is MCPOrb different from traditional MCP servers?

Traditional MCP servers require continuous running and network exposure. MCPOrb bundles knowledge, index, and runtime into a single portable file that runs locally with zero data leakage. No cloud dependency, no always-on server needed.

Which AI editors support MCPOrb?

MCPOrb natively supports Claude Desktop, Cursor, VS Code, and any MCP-compatible client. It exposes a standard search_knowledge tool directly to your AI assistant.

What is Model Context Protocol (MCP) and how does MCPOrb leverage it?

MCP is an open standard enabling AI clients to securely access data sources. MCPOrb implements this by packaging documentation into a single, portable, local file exposing a native search_knowledge tool to LLM clients like Cursor or Claude Desktop.

How does MCPOrb support Harness Engineering and AI-native architecture?

MCPOrb provides the structural framework for Harness Engineering, shifting AI value from raw model capability to deterministic, structured system architecture. It enables ANTcell setups — solo engineers distributing secure, mathematically certain knowledge packs without external SaaS leak risks.

How to get a free trial of MCPOrb Builder?

Register an account to receive a free 30-day trial code. Upload an Orb asset bundle to extend it to a 90-day free trial of the Builder. See the upload page for details.

Can MCPOrb run on Linux or older hardware?

Yes. The MCPOrb runtime is lightweight and cross-platform — Apple Silicon (M1/M2/M3), Linux (Ubuntu 20.04+, Debian), Windows x64. It runs on older hardware with zero cloud dependencies or external database APIs.

Found an issue?

Report a problem or suggest an improvement.

Help us make MCPOrbStore better. Whether it's a bug, a missing feature, a license issue, or something else — let us know.

Fast responseWe review reports within 1 business day
TrackedEvery report is reviewed by the team
No account neededSubmit anonymously or leave your email