M
MCP Directory CURATED DIRECTORY
911 MCP servers indexed
中文
← Back to directory
N

Neo.mjs Agent OS

Community
Framework / SDK
An AI engineering team inhabiting live applications

This is a framework for building MCP servers or agents, not a server you connect a client to.

How to install ↓ Compare GitHub source repository ↗
Category
Dev Tools #394 of 438
Stars
★ 3.3k Very popular
Runtime
Node.js
Credentials
No credential needed
License
MIT
Last commit
34FMRS · D

Based on the supplied README, Neo.mjs is an open-source project combining an Agent OS with a multi-threaded frontend runtime. The README explicitly describes several MCP servers and their roles, but provides no concrete server manifest, tool identifiers, transport, client compatibility, or installation configuration. Those fields therefore remain empty.

Strongest · Maintenance 13/20 Weakest · Reliability 3/20

Reliability
3/20
Security and permissions
6/20
Maintenance
13/20
Documentation
8/20
Setup experience
4/20
Why each score
Reliability 3/20
No server manifest, source code, launch command, tool definitions, tests, or CI evidence is provided. The README only claims that several MCP servers exist, so MCP initialization, tool-list handshake, dependency control, and error behavior cannot be verified. The score is heavily reduced for missing executable evidence.
Security and permissions 6/20
The README mentions sandboxed file I/O, an OIDC ingress, and tenant isolation, but provides no reviewable evidence for permission scopes, credential handling, network access, data flows, dangerous-operation confirmation, or implementation. GitHub workflows, runtime mutation, and cloud deployment imply meaningful side effects. There is no evidence of malware or credential theft, so no red line is triggered, but controls remain insufficiently auditable.
Maintenance 13/20
The repository is marked non-archived, uses the MIT license, and describes named human and AI maintainers, ongoing PR/issue activity, and a public contribution workflow. These claims are primarily README evidence; verifiable release cadence, commit history, dependency updates, and security-response procedures are absent, so the score cannot approach full marks.
Documentation 8/20
The README gives extensive conceptual and architectural coverage and links to deployment, architecture, and Agent OS material. However, it lacks a reviewable MCP manifest, tool parameters, authentication configuration, limits, cost information, troubleshooting, and client connection examples. Conceptual documentation does not substitute for operational server documentation.
Setup experience 4/20
The only concrete command is npx neo-app@latest for bootstrapping an application; it does not establish MCP server installation or client connectivity. Runtime requirements, server startup configuration, mainstream client examples, authentication steps, and platform compatibility evidence are missing, making the setup path largely unverifiable.

Static review · not runListed 2026-07-29

Read the FMRS scoring method →

Fit and risk

What it can accessReads local filesWrites / deletes local filesUses the networkConnects to a database

Best for

  • AI architects needing persistent memory and cross-model coordination
  • Engineers building enterprise multi-window applications, trading platforms, or IDE-class tools
  • Researchers studying self-evolving systems, gated recursive self-improvement, and multi-agent governance

Not for

  • Static content sites or simple blogs
  • Teams seeking a drop-in syntax replacement
  • Developers unwilling to adopt the Worker Actor Model or treat AI as a peer maintainer

Required permissions

  • Access to source code, documentation, issues, pull requests, and discussions for knowledge-base construction
  • Persistent storage for agent memory, messages, and graph data
  • If GitHub Workflow is enabled, management of issues, pull requests, reviews, labels, and projects
  • If Neural Link is enabled, inspection of live application state and mutation of the running application
  • Sandboxed file access may be used by internal Neo.ai.Agent local loops

Risks and side effects

  • Neural Link can modify live application state and hot-patch code, so incorrect operations may affect running applications
  • GitHub workflows may create or review pull requests, manage issues, and synchronize repository state
  • Persistent memory and knowledge stores may retain repository content, discussions, and agent reasoning
  • Multi-tenant deployments must preserve tenant identity and visibility isolation; the README gives no specific security configuration

Setup

Before you start

Runtime：Node.js

The source does not provide an MCP server manifest, launch command, transport, or client configuration example, so verifiable installation steps cannot be supplied. The README provides npx neo-app@latest for creating a Neo.mjs application workspace, but does not state that this installs the MCP server.

Check that it works

No tools or client config were recorded in the listing; confirm setup by running npx neo-app@latest and checking, per the linked Deploying the Agent OS docs, that the Knowledge Base / Memory Core MCP servers appear in your client's tool list.

Troubleshooting

  1. Confirm that the repository root is used, because the server path is specified as the repository root
  2. Check the client configuration; the source provides no usable server manifest or installation configuration
  3. Confirm that the required Knowledge Base, Memory Core, GitHub Workflow, or Neural Link components are deployed
  4. For GitHub operations, check repository access and final merge authority
  5. For live application operations, confirm that the target application is running and accessible through Neural Link

Things to try

Once connected, you can ask your AI assistant things like:

  • Use the Knowledge Base MCP server to answer questions about this repository's architecture.
  • Query the Memory Core MCP server for the Native Edge Graph's current Golden Path priorities.
  • Point the Agent OS at my repository and onboard it as a tenant.
  • Use the Neural Link MCP server's get_component_tree to inspect the live App Worker state.

Use cases

Build semantic knowledge and persistent memory for a codebase
Coordinate multiple models for engineering tasks, code review, and GitHub workflows
Inspect and mutate a running Neo.mjs application
Study multi-agent systems, self-evolution, and runtime embodiment

Overview

Neo.mjs is an open-source project described as a self-evolving software organism. It combines a cross-model AI engineering team, persistent memory, Active Hybrid GraphRAG, DreamService, self-healing loops, and Neural Link. Its Brain (/ai/) provides the Agent OS, while its Body (/src/) is a multi-threaded frontend application runtime. The README describes Knowledge Base, Memory Core, GitHub Workflow, Neural Link, and File System MCP servers, but no server manifest is available.

Similar servers

Context7 80 · B

Upstash's official server providing up-to-date third-party library docs for AI coding assistants

★ 63.3k · Tools 2 Compare with this →
Compare with other MCP servers

Source revision a4eafc176c50 Data synced 2026-10-19 Read the FMRS scoring method