AgentIndex icon
AgentIndex
ToolsMCP ServersHOTCategoriesTrendingNewBest ListsCompareGuides
Submit Tool
ToolsMCP ServersHOTCategoriesTrendingNewBest ListsCompareGuides
Home/
Security & Safety/
toolhive
toolhive logo

toolhive

Active·★ 2.3k·Apache-2.0·Updated 2026-10-27
★ Essential★ Trending

ToolHive is a platform for running MCP servers in production. It has four parts. The Runtime deploys servers in containers via Docker or Podman locally, or on Kubernetes with an Operator, and can proxy remote MCP servers, with OpenTelemetry and Prometheus for monitoring and audit logging. The Registry Server holds a curated catalog of trusted servers, integrates with the official MCP registry and can verify provenance and signatures. The Gateway defines endpoints and access policies, supports SSO through OIDC/OAuth, can combine tools into a virtual MCP with a deterministic workflow engine and can filter tool descriptions to save tokens. The Portal is a cross-platform desktop app and browser UI for one-click installs.

The README stresses that each server runs in an isolated container with only the permissions it needs and that secrets are managed rather than stored in plaintext. Individual developers can start with the desktop UI or CLI, which can run any MCP server from a container image or build one from common package managers, manage encrypted secrets, control network isolation and test with the official MCP Inspector, then use the same ideas in the Kubernetes Operator. It auto-configures clients such as GitHub Copilot and Cursor. Quickstarts for the desktop app, CLI and Operator are at docs.stacklok.com, and the project is Apache licensed.

This write-up is based on the project's README and public repository data. We have not hand-tested the tool. Data as of 2026-10-25.

License
Apache-2.0
Latest release
v0.56.0 · 2026-10-22
Last commit
2026-10-24
Contributors
138
Forks
323
Open issues and PRs
402
First published
2025-03-12

Source: GitHub, npm and PyPI public APIs.

#Model Context Protocol#Server Management#Security#Deployment#Kubernetes
$ Install
$ git clone https://github.com/stacklok/toolhive.git
↗ Visit site★ GitHub
01

Features

01Instant deployment with one click or command using Docker or Kubernetes.
02Secure by default, running servers in isolated containers with managed secrets.
03Works everywhere, available as a desktop app, web app, CLI, and Kubernetes Operator.
04Seamless integration with popular AI clients like GitHub Copilot, Cursor, and VS Code.
05Modular architecture comprising Gateway, Registry Server, Runtime, and Portal for comprehensive management.
02

Why choose it

+Servers run in isolated containers with scoped permissions and managed secrets
+Same tooling from a laptop (desktop app or CLI) to a Kubernetes Operator
+Registry with provenance verification, SSO-capable gateway and audit logging for teams
+Auto-configures popular AI clients and includes the MCP Inspector
03

Trade-offs

!Needs Docker, Podman or Kubernetes; heavier than just running an MCP server directly
!Four components (Runtime, Registry, Gateway, Portal) take time to understand for small teams
!Several parts live in separate repositories, so you follow multiple docs
!Container isolation reduces risk but does not replace reviewing what each server can reach
05

Maintenance check

ActiveChecked: 2026-10-27
Last commit
2026-10-27
Commits, last 90 days
100+
Releases, last 90 days
25
Latest release
v0.58.0 (2026-10-27)
Open issues
398
Recent issues with at least one comment
4/17
Commits per week, last 12 weeks
Commits per week, last 12 weeks: 46, 40, 33, 30, 113, 35, 25, 10, 12, 44, 56, 24

Oldest week on the left; the latest four weeks are orange. Source: GitHub commit activity.

Status rule: archived on GitHub = Archived; last commit within 90 days = Active; 91 to 365 days = Slowing; older = Stale. All figures come from the GitHub API on the date shown.

Most discussed open issues

  • #5860 vmcp: all VirtualMCPServer instances grow 2-3x in memory over ~24h regardless of client activity14 comments · 2026-07-17
  • #4048 Distinguish tool discovery from tool invocation in authorization and audit12 comments · 2026-03-09
  • #5321 Embedded auth server rejects DCR from confidential MCP clients (`token_endpoint_auth_method` hardcoded to `none`)9 comments · 2026-05-19
  • #6831 Re-authorization walks the full upstream chain instead of evaluating existing credentials8 comments · 2026-10-15
  • #6675 Add a group-aware UpstreamFilter implementation for the embedded auth server8 comments · 2026-09-17

Issue titles are quoted from GitHub. Open the link to check whether a problem still applies to the version you use.

04

Who it fits

Pick it if
  • You want MCP servers to run in isolated containers with scoped permissions and managed secrets, on Docker, Podman or Kubernetes.
  • You want the same tooling from a laptop (desktop app or CLI) to a Kubernetes Operator.
  • Your team needs a registry with provenance verification, an SSO-capable gateway and audit logging.
Look elsewhere if
  • You just want to run one MCP server: it needs Docker, Podman or Kubernetes and is heavier than starting the server directly.
  • You want a single component: the Runtime, Registry, Gateway and Portal take time to understand, and parts live in separate repositories.
  • You expect container isolation to replace reviewing what each server can reach.
06

How it differs

archestra logovs archestra★ 4.4k

ToolHive is a runtime and registry for MCP servers, under an Apache licence. Archestra is a broader enterprise platform with chat, LLM and MCP gateways, and dual AGPL and commercial licensing.

Full comparison →
07

Compatibility

GitHub Copilot
Supported Client
Verified via docs
Cursor
Supported Client
Verified via docs
Claude Desktop
Supported Client
Verified via docs
VS Code
Supported Client
Verified via docs
Docker
Native Runtime
Verified via docs
Podman
Native Runtime
Verified via docs
08

Quick start

1
$ git clone https://github.com/stacklok/toolhive.git
09

Use cases

↳Individual developers quickly deploying and testing MCP servers locally using the desktop UI or CLI.
↳Organizations centrally managing and orchestrating a fleet of MCP servers in Kubernetes for production and scale.
↳Enterprises building hybrid solutions, integrating MCP servers into custom internal tools and agentic workflows with unified access controls.
10

Alternatives

⚡Top VS:
VSarchestra★4.4k→VSGhidraMCP★11.6k→VSokta-mcp-server★38→VSoctocode-mcp★956→
archestra logo
archestra★ 4.4k
Archestra is an open-source enterprise AI platform: an internal chat assistant, LLM gateway, MCP gateway and registry, A2A gateway and agent runtime, built on SSO, RBAC, sandboxed execution, guardrails and OpenTelemetry.
VSCompare→
GhidraMCP logo
GhidraMCP★ 11.6k
MCP Server for Ghidra
VSCompare→
okta-mcp-server logo
okta-mcp-server★ 38
The Okta MCP Server is a groundbreaking tool built by the team at Fctr that enables AI models to interact directly with your Okta environment using the Model Context Protocol (MCP). Built specifically for IAM engineers, security teams, and Okta administrators, it implements the MCP specification to help work with Okta enitities
VSCompare→
octocode-mcp logo
octocode-mcp★ 956
MCP server for semantic code research and context generation on real-time using LLM patterns | Search naturally across public & private repos based on your permissions | Transform any accessible codebase/s into AI-optimized knowledge on simple and complex flows | Find real implementations and live docs from anywhere
VSCompare→
mcp-for-security logo
mcp-for-security★ 628
MCP for Security: A collection of Model Context Protocol servers for popular security tools like SQLMap, FFUF, NMAP, Masscan and more. Integrate security testing and penetration testing into AI workflows.
VSCompare→
ADR logo
ADR★ 2.1k
ADR (Agentic Detection and Response): open-source telemetry sensor and dual-agent detection framework for securing MCP-based AI agents.
VSCompare→
ida-pro-mcp logo
ida-pro-mcp★ 13.2k
AI-powered reverse engineering assistant that bridges IDA Pro with language models through MCP.
VSCompare→
hexstrike-ai logo
hexstrike-ai★ 13.1k
HexStrike AI MCP Agents is an advanced MCP server that lets AI agents (Claude, GPT, Copilot, etc.) autonomously run 150+ cybersecurity tools for automated pentesting, vulnerability discovery, bug bounty automation, and security research. Seamlessly bridge LLMs with real-world offensive security capabilities.
VSCompare→
See all alternatives →

Featured in these lists

  • MCP Gateways and Proxies Compared
  • Open-Source Security Tools for MCP Servers and AI Agents

Comments

Log in to leave a comment

No comments yet. Be the first!

On this page
01Features02Why choose it03Trade-offs05Maintenance check04Who it fits06How it differs07Compatibility08Quick start09Use cases10Alternatives
stacklok /
toolhive
↗

Run and govern MCP servers in locked-down containers, locally or on Kubernetes

2.3k323Apache-2.0
Last commit1d ago
StatusActive
CategorySecurity & Safety
Contributors138
Latest releasev0.56.0
Links
Documentation↗Discussion↗Issues↗Releases↗
Cloud Infrastructure
Ad

Deploy on DigitalOcean

Host your MCP servers, Docker containers, and AI agents with $200 free credit over 60 days.

Claim $200 Free Credit
© 2026 AgentIndex.app|Built by a 10-year iOS Developer.
Best by use caseAboutContactAdvertisePrivacy PolicyTerms of UseQYSGitHubBuy me a coffee ☕

Browse by Category

MCP ServersCode AssistantWorkflow AutomationRAG / Knowledge BaseMulti-AgentBrowser AutomationLLM InfraDev ToolingObservability

Not affiliated with Anthropic, OpenAI or Microsoft.