Model Context Protocol server

Every AI session at iamota
starts knowing where it works.

This server hands Claude the context an iamota project needs before you type anything — who we are, how your department works, the coding standards for the repo you’re in, and what has actually been agreed on the project. One source, every tool, no pasting briefs into chat windows.

Endpoint
https://mcp.iamota.com/mcp
Sign in
Google, @iamota.com only
Access
Scoped to your role groups

What it is

Institutional knowledge, as a protocol.

iamota’s standards, skills, templates and project briefs live as markdown in a single repository. This server indexes that repository and serves it over the Model Context Protocol, so any MCP-capable assistant can query it live instead of relying on whatever someone last pasted into a prompt.

Context arrives in layers. Org context establishes who iamota is and how we talk about our work. Department context adds how your discipline operates. Workspace standards load per repository type. Project context loads per client engagement. You get the layers your role grants, and nothing else.

  1. 01

    You open a session

    In Claude, Claude Code, or any connected client.

  2. 02

    Context loads itself

    One call returns org, department, workspace and project context together.

  3. 03

    Detail on demand

    Full standards docs, templates and project files are pulled only when needed.

What it serves

Five kinds of content.

Access is fail-closed and driven by your Google Workspace group membership. Content you are not cleared for is invisible, not merely hidden.

Get started

Pick where you work.

Connecting takes a couple of minutes. You will sign in with Google using your iamota account — there are no keys or tokens to manage.

Before you connect

Your access comes from Google Workspace groups named mcp-<role>@iamota.com — one for your department, plus any overlays that describe you. If you have just joined, ask your team lead to add you. Connecting without a role group works, but you will see very little.

If something is missing

Gaps are usually a markdown fix.

If Claude answers wrong, misses context you expected, or you hit an access error, say so in the AI channel in Slack or tell Steve Reimer. Most gaps are a small content change in the source repository, and the server redeploys on every merge.

Curious what you have access to? The docs app browses every document the server holds for your roles, and its simulator shows exactly what a session start returns before you spend a single token on it.