MCP.Market

Zoogent MCP server

by bepunk·io.github.bepunk/zoogent·v0.4.2·3 stars
No reviews yet

AI agent orchestrator with Architect AI. Multi-team, web dashboard, 35+ MCP tools.

C57/100grade C
Write a reviewInstall
What users say
No reviews yet
Be the first
Safety scan
C57/100

full report

Adoption
Growing

3 stars95 downloads/wk

Reviews

Write one

Nobody has reviewed Zoogent yet.

If you have run it, two minutes of your experience saves the next person an afternoon.

Zoogent tools (37, 13 write)

write = sends, deletes, buys or posts

Read from the package source without running it. The installed server may list more.

  • add_memory

    Add a memory entry for an agent

  • approve_knowledge

    Approve a draft team knowledge entry

  • archive_knowledge

    Archive a team knowledge entry

  • assign_skill

    Assign a skill to an agent

  • create_agentwrite action

    Register a new agent. Default runtime is "typescript": provide `source` with TS code (zoogent bundles and stores it; the agent is ready to run). Omit source to register the agent and upload code later via write_agent_code. For binaries or non-TS scripts use runtime="exec" with `command` and `args`. After create: assign_skill, then trigger_agent to test.

  • create_integrationwrite action

    Add an integration (3rd party API key) to an agent. Env vars are injected on spawn as INTEGRATION_{NAME}_{FIELD}.

  • create_skillwrite action

    Create a new skill in the selected team

  • create_taskwrite action

    Create a task for an agent

  • create_teamwrite action

    Create a new team and auto-select it

  • delete_agentwrite action

    Remove an agent and its code

  • delete_integrationwrite action

    Remove an integration from an agent

  • delete_memorywrite action

    Delete a memory entry

  • disable_agent

    Disable an agent and unschedule

  • enable_agent

    Enable an agent and reschedule

  • get_agent

    Get agent details including runs, skills, and bundle status

  • get_agent_code

    Read the current TypeScript source and bundle status for an agent. Use to inspect before editing.

  • get_agent_guide

    Load design methodology and reference docs. Topics: team-design (Jobs-Roles-Flows), agent-patterns (boilerplate examples), code-generation (blessed deps, SDK API, import examples), skill-writing, debugging, platform-rules. Omit topic to read all. Call before designing agents.

  • get_budget_status

    Get spending vs budget for all agents

  • get_costs

    Get cost summary by agent and model

  • get_logs

    Get stdout/stderr for an agent run

  • get_memories

    Get agent memories, optionally filtered by tags or search

  • get_skill

    Get skill content and metadata

  • get_started

    Check ZooGent status, list teams, and guide the user through setup. Call this first.

  • get_task

    Get task details including evaluations

  • list_agents

    List all agents in the selected team with status, last run, monthly cost, and bundle readiness

  • list_integrations

    List integrations for an agent (credentials masked)

  • list_skills

    List all skills in the selected team

  • list_tasks

    List tasks, optionally filtered

  • list_team_knowledge

    List team knowledge entries

  • list_teams

    List all teams

  • select_team

    Select a team to work with. All subsequent operations will be scoped to this team.

  • trigger_agent

    Manually trigger an agent run. Requires the agent to have runnable code (typescript: uploaded source; exec: command set).

  • unassign_skillwrite action

    Remove a skill from an agent

  • update_agentwrite action

    Update agent configuration (not code — use write_agent_code for source).

  • update_memorywrite action

    Update an existing memory

  • update_skillwrite action

    Update skill content

  • write_agent_codewrite action

    Upload TypeScript source for a typescript-runtime agent. Zoogent bundles with esbuild and replaces the stored code atomically. Returns esbuild errors on failure (e.g. unknown import — not in the blessed deps list). Does not apply to exec-runtime agents. Use to iterate after create_agent.

Public scan report

scanner v0.1.9 · 2026-09-27 · same rubric, same numbers if you re-run it

2 medium
  • Code scan115 source files scanned15/25
  • –Live reliabilityno gateway calls yet and no remote to proben/a
  • –Tool poisoningtools not inspected (local package is not executed); not countedn/a
  • Auth qualitystatic API keys via environment variables6/15
  • Maintenancelast push 144 days ago8/15
  • Maintainer identityregistry namespace matches repository owner; GitHub account older than a year8/10

Findings (2)

  • mediumnpm install lifecycle script presentinstall.script
  • mediumnpm install lifecycle script presentinstall.script
    package.json: … "prepublishOnly": "npm run build", "postinstall": "node -e \"console.log('\\n ZooGent in…
Overall 57/100. Components that don't apply are left out of the denominator. Any critical finding is an F.RubricAppeal a findingJSON

How Zoogent compares

Against every graded AI and agents MCP server, from our own scans and uptime checks. Updated after each scan run.

Safety score
57/100
higher than 14% of 4,460 AI and agents servers
Tools
37
the typical AI and agents server has 6
Tools that change things
35%
typical: 0%. Fewer write tools means less an agent can break
GitHub stars
3
more than 75% of 2,711 AI and agents servers
Downloads a week
95
more than 63% of 1,538 AI and agents servers with a package

What its tools let an agent do: 17 read · 7 create · 8 change · 4 delete · 1 run or deploy

Install Zoogent in Claude Code, Cursor or VS Code

Runs npx -y zoogent on your machine. Read the scan report first; the gateway never runs local packages.

claude mcp add zoogent -- npx -y zoogent
Add to Cursor

What the publisher says

From the Zoogent repository's README, as published. We do not edit it. Read it on GitHub

<!-- mcp-name: io.github.bepunk/zoogent -->

ZooGent

Lightweight AI agent orchestrator with built-in Architect AI. Multi-team support - run isolated agent teams in a single instance. Describe what you want to automate, get working agents.

Quick Start

npx zoogent create my-agents
cd my-agents
npx zoogent start

Open http://localhost:3200. Create account > create team > add Anthropic API key in Team Settings > go to Architect and describe what you want to automate.

What is ZooGent

ZooGent is a process manager for AI agent teams. It spawns agents, routes tasks between them, tracks costs, and captures logs. Each agent is a standalone script that calls its own LLM.

Multi-team isolation - one instance, multiple teams. Each team has its own agents, skills, memory, knowledge, Architect chat, and API keys. Teams don't see each other's data.

Two ways to use it:

  1. Chat UI - open the Architect page in your browser, describe your task in plain language. The Architect AI designs the team, creates skills, writes agent code, and tests everything.
  1. Claude Code + MCP - connect MCP to Claude Code, build agents from the terminal with full control over code and configuration.

Both paths use the same API, same database, same agents. Pick whichever fits your workflow.

Getting Started

Path 1: Web UI + Architect

Local

npx zoogent create my-agents
cd my-agents
npx zoogent start
  1. Open http://localhost:3200
  2. Create your account
  3. Create a team
  4. Go to Team Settings > add your Anthropic API key
  5. Go to Architect > describe what you want to automate

Architect designs agents, writes skills, generates code, and tests everything through conversation.

Server

npx zoogent create my-agents
cd my-agents
npx zoogent start -d

Set BETTERAUTHURL to your public URL in .env. Use a reverse proxy (nginx, Caddy) or deploy via Docker (see Deployment section). The web UI is the same - just accessed remotely.

Path 2: Claude Code + MCP

Local

Start the server locally, then connect Claude Code via MCP:

npx zoogent create my-agents
cd my-agents
npx zoogent start -d

Add the MCP server to Claude Code. Run this from inside the project you want to work in — each project typically binds to its own ZooGent instance (its own SQLite, teams, agents, API keys), so the MCP config should live with the project:

claude mcp add zoogent -s project -- npx zoogent mcp

-s project writes the config to .mcp.json in the project root — commit it and teammates who clone the repo get the same MCP setup automatically.

Claude Code auto-discovers the local server. Ask Claude to create a team and design your agents.

Remote server

Deploy ZooGent to a server (see Deployment). Open the web UI, create an account, go to Settings > generate an API key.

Run from inside the project that should connect to this ZooGent instance:

claude mcp add zoogent -s project \
  -e ZOOGENT_URL=https://your-domain.com \
  -e ZOOGENT_API_KEY=zg_your-key-from-settings \
  -- npx zoogent mcp

Each project can point at its own ZooGent (local or remote), with its own URL and API key. Keeping the config in the project's .mcp.json makes that mapping explicit. Use -s user instead only if you have a single shared ZooGent instance across all projects.

Alternative: configure via .mcp.json

{
  "mcpServers": {
    "zoogent": {
      "command": "npx",
      "args": ["zoogent", "mcp"],
      "env": {
        "ZOOGENT_URL": "https://your-domain.com",
        "ZOOGENT_API_KEY": "zg_your-key-from-settings"
      }
    }
  }
}

Claude Code connects to the remote server. Create teams, design agents, write code - all through MCP tools. Agents run on the server.

How It Works

Shortened. The full README is on GitHub.

Nothing above is checked by us. What we check is on the safety report.

Zoogent: common questions

Is Zoogent MCP server safe?
With care: it is graded C, so read the findings first (57/100). Read the Zoogent safety report
How do I install Zoogent?
It runs on your machine. Copy the Claude Code, Cursor, VS Code or Claude Desktop config from the install section.
Does Zoogent need an API key?
Yes. The registry entry asks for ZOOGENT_API_KEY.
Is Zoogent maintained?
The last commit was 182 days ago (2026-05-06). The latest release is v0.4.2.
What can I use instead of Zoogent?
Servers from other publishers that do the same job: agent-device MCP server, PaperBanana MCP server and You.com Web Access & AI MCP server.
How does Zoogent compare with other AI and agents MCP servers?
safety score 57/100 (higher than 14% of 4,460 AI and agents servers); tools 37 (the typical AI and agents server has 6); tools that change things 35% (typical: 0%. Fewer write tools means less an agent can break). Full comparison.

Alternatives to Zoogent

Same job from other publishers: the closest match first, then the best rated.

  • agent-device
    Open source device layer for AI agents: control mobile, web, TV, and desktop apps over MCP
    A
  • PaperBanana
    Generate academic diagrams and statistical plots from text using multi-agent AI.
    A
  • You.com Web Access & AI
    Web search, AI agent, and content extraction via You.com APIs
    A
  • Robin: AI-Powered Dark Web OSINT
    Dark web OSINT over Tor: search onion engines, scrape pages, report with your own model.
    A
  • three.ws Brain
    List LLM providers and run chat completions through the three.ws multi-provider router.
    B