๐ BIG NEWS ๐ โฆ Avo 4 has officially SHIPPED โฆ ๐ the beta is over, 4.0 is here โฆ ๐ ๏ธ build admin panels, dashboards & internal tools at light speed โฆ โญ you are visitor #1,137,351 โฆ ๐ tell a friend โฆ best viewed in Ruby on Rails โฆ ๐ click here to see what’s new โฆ
MCP
coming soonConnect your admin to AI agents over the Model Context Protocol, so assistants can safely read and act on your data.
# $30/mo
gem "avo-mcp"
Let Claude run your admin. MCP turns your Avo app into a Model Context Protocol server, so an AI assistant can discover your resources, search and inspect records, create and update data, and fire your Avo actions, all from a chat prompt. Ask "find the users with admin in their email" or "create an article titled Getting Started by Sarah Chen" and the model calls the right tools against your real resources.
It exposes nine tools (from list_resources and
search_records through create_record,
run_action, and delete_record) and connects over
stdio for Claude Desktop and Claude Code, or as a token-authenticated HTTP
endpoint mounted in your Rails app. Every call runs through your existing Avo
authorization, so the model can only do what the current user is allowed to
do. Building a safe AI bridge to your admin yourself means writing and
auditing that whole tool layer; here you add a gem instead.
See it in action
What you get
-
Nine MCP tools:
list_resources,list_records,show_record,search_records,create_record,update_record,delete_record,list_actions, andrun_action - Stdio transport for Claude Desktop and Claude Code, plus a token-authenticated HTTP endpoint at a configurable mount path
- Every call runs through your existing Avo policies, so the model can only read or change what the current user may
-
A
current_userproc resolves the acting user per request, scoping authorization to whoever the chat is acting as - Runs your custom Avo actions on a record or a batch, the same actions you already defined in the admin
-
Structured error types (
invalid_params,not_found,not_authorized,validation_error) so the model gets a clear answer back
Why it pays off
- An AI bridge to your admin is mostly the tool layer and its guardrails: ship that this week instead of speccing it over a sprint.
- Authorization is the hard part to get right. This reuses your Avo policies instead of you re-implementing access control for a second surface, where one missed check hands the model write access.
- Every Avo release makes it better: the MCP server, its transports (stdio and HTTP), and the exposed tools stay aligned with your resources and keep working through upgrades, without you patching them.
- One less internal service for your team to build, secure, document, and support.