Hq MCP server
{
"mcpServers": {
"giljo-hq": {
"type": "http",
"url": "https://app.giljo.ai/mcp"
}
}
}{
"mcpServers": {
"giljo-hq": {
"type": "http",
"url": "https://app.giljo.ai/mcp"
}
}
}{
"mcpServers": {
"giljo-hq": {
"type": "http",
"url": "https://app.giljo.ai/mcp"
}
}
}{
"servers": {
"giljo-hq": {
"type": "http",
"url": "https://app.giljo.ai/mcp"
}
}
}{
"mcpServers": {
"giljo-hq": {
"type": "http",
"url": "https://app.giljo.ai/mcp"
}
}
}file .mcp.json (in your project root)file claude_desktop_config.jsonfile ~/.cursor/mcp.jsonfile .vscode/mcp.jsonfile ~/.codeium/windsurf/mcp_config.jsonversion 2.0.4
What it does
Giljo HQ is a self-hosted MCP server that stores product context, project definitions, and accumulated development memory, then assembles structured prompts for AI coding agents to consume. It coordinates multi-agent workflows and tracks job progress in real time without calling any AI model itself. It is aimed at developers and product managers who want a persistent, shared knowledge base across multiple AI coding tool sessions.
- Store a product's tech stack, architecture, and vision document so every connected AI coding agent session starts with consistent context.
- Assemble and stage a bootstrap prompt for a project, then monitor connected agents' real-time status, step progress, and messages from a dashboard.
- Search and retrieve accumulated 360 Memory entries from previous completed projects so later agent sessions inherit decisions and patterns already discovered.
Tools this server exposes
| Tool | Does |
|---|---|
| giljo_setup | Installs or refreshes agent templates on the connected platform. |
| giljo | Slash/dollar command skill for capturing tasks, creating or updating projects, and looking them up mid-session from a CLI tool. |
Before you enable it
An MCP server runs with whatever access you give it. Start with a development account rather than production, and check which tools the server exposes before you let an agent call them unattended.