Gluesync Module

Core Hub MCP server

An embedded Model Context Protocol server runs inside Core Hub and exposes pipeline management, debugging, observability, connections, and SQL tasks as tools an AI agent can call — with the caller’s own token and the same RBAC as the Web UI.

MCP client connecting to Gluesync Core Hub
  1. 01 Connect an MCP client
  2. 02 Call RBAC-safe tools
  3. 03 Query via Query Studio tools

Why Core Hub MCP

Operate Core Hub from the agents you already use

01 / Sidecars

Extra MCP proxies

Extra MCP proxies mean another process to secure and upgrade.

02 / Privilege drift

Keep the audit boundary

Agent credentials that bypass Control Plane roles break audit.

03 / Outcome

Use any MCP client

Point Claude, Cursor, MCP Inspector, or any MCP client at Core Hub and keep the same permissions as the UI.

How it works

Same JVM, same RBAC, standard MCP transports

  1. 01

    Embed

    Mount beside REST

    corehub-mcp mounts /mcp (SSE) and /mcp/http (streamable HTTP) beside REST.

  2. 02

    Authenticate

    Use the caller's identity

    Bearer session JWT or Personal API Token (gsp_…); zero elevated privileges in the module; loopback REST into Core Hub.

  3. 03

    Act

    Call governed tools

    125 built-in tools in Gluesync 2.3 (Beta) plus optional per-session Query Studio database tools when enabled.

Capabilities

Core Hub operations exposed as governed MCP tools

The Gluesync 2.3 (Beta) catalog is 125 built-in tools (73 read and 52 write), including Chronos schedules, Core Hub webhooks, email notifications, and the AI agents category that lets any MCP client start published Spark agents and search the Enterprise brain. Automator/Bootstrapper MCP remains a separate local/stdio path.

Built into Core Hub

125 built-in tools in Gluesync 2.3 (Beta)

Operational coverage without a separate proxy

Use one embedded server for Core Hub management, diagnostics, reporting, and data tasks.

  • Create, assign, and configure pipelines — not only operate existing ones
  • Pipelines, agents, entities, sync, notifications, metrics, Chronos, Validator, Core Hub webhooks, and email
  • Mapping functions, Query Studio, documentation search, groups, configuration, license, version, export, and Connections
  • AI agents: list_published_agents, start_agent_run, get_agent_run, publish_skill, and search_enterprise_brain for Claude, ChatGPT, Grok, Meta Muse, and any MCP client

View Core Hub MCP docs ↗

MCP connectivity

Standard transports

SSE and streamable HTTP

Connect a compatible client through SSE at /mcp or streamable HTTP at /mcp/http.

  • Streamable HTTP example: https://corehub-host.example/mcp/http
  • Authorization: Bearer YOUR_TOKEN (session JWT or gsp_ personal API token)
  • SSE endpoint: /mcp

View connection guidance ↗

Control Plane security

Governance

The caller's token and permissions stay in force

The module adds zero elevated privileges and routes operations back into Core Hub under the authenticated caller.

  • Read-only mode, allow-lists, and on-by-default controls
  • No extra privileges beyond the caller's token and Control Plane RBAC
  • Tool exposure can be bounded to the operations you choose

View governance controls ↗

Resources

Core Hub MCP documentation and resources

Roadmap

Public Roadmap

What we're building next. Submit feature requests.

View roadmap ↗
Support

Support

Access technical support and operational assistance.

Get support ↗
Status

Service Status

Real-time platform availability and incident history.

Check status ↗

Operate Gluesync from your MCP client

Use the embedded server with the same bearer identity and Control Plane permissions as the Web UI.