Extract the dashboard into the actionagent gem, with feature parity

.gitignore

@@ -6,6 +6,7 @@
 /doc/
 /log/*.log
 /node_modules/
+node_modules/
 /pkg/
 /test/dummy/db/*.sqlite3
 /test/dummy/db/*.sqlite3-*

AGENTS.md

@@ -11,6 +11,8 @@
 | Provider implementations | `lib/active_agent/providers/` |
 | Agent concerns/mixins | `lib/active_agent/concerns/` |
 | Rails generators | `lib/generators/active_agent/` |
+| Dashboard engine (`actionagent` gem) | `actionagent/` |
+| Dashboard React source | `actionagent/frontend/` |
 | Test suite | `test/` |
 | Test Rails app | `test/dummy/` |
 | Documentation source | `docs/` |
@@ -79,8 +81,11 @@ end
 
 ### Template Structure
 
-Templates live in `app/views/agents/{agent_name}/`:
-- `instructions.md.erb` - System prompt (shared across actions)
+Templates live in `app/views/agents/{name}/`, where `{name}` is the class name
+underscored with the `_agent` suffix dropped — `TravelAgent` looks in
+`app/views/agents/travel/` (`_prefixes` in `lib/active_agent/concerns/view.rb`
+also searches the full underscored class name, `app/views/travel_agent/`):
+- `instructions.md` - System prompt (shared across actions); `.md.erb` works too
 - `{action_name}.md.erb` - Action-specific prompt template
 
 ### Provider Configuration
@@ -102,11 +107,13 @@ development:
 rails generate active_agent:agent AgentName action1 action2
 ```
 
-Creates:
+Creates (the view directory drops the `_agent` suffix, the agent file keeps it):
 - `app/agents/agent_name_agent.rb`
-- `app/views/agents/agent_name_agent/instructions.md.erb`
-- `app/views/agents/agent_name_agent/action1.md.erb`
-- `app/views/agents/agent_name_agent/action2.md.erb`
+- `app/views/agents/agent_name/instructions.md`
+- `app/views/agents/agent_name/action1.md.erb`
+- `app/views/agents/agent_name/action2.md.erb`
+- `test/agents/agent_name_agent_test.rb` and
+  `test/docs/previews/agent_name_agent_preview.rb`
 
 ### Adding a Tool to an Agent
 
@@ -167,12 +174,28 @@ end
 
 ### Adding a New Provider
 
-1. Create `lib/active_agent/providers/my_provider.rb`
-2. Create `lib/active_agent/providers/my_provider/` directory with:
-   - `client.rb` - API client wrapper
-   - `request.rb` - Request building
-   - `response.rb` - Response parsing
-3. Register in `lib/active_agent/providers.rb`
+1. Create `lib/active_agent/providers/{service}_provider.rb` defining
+   `ActiveAgent::Providers::{Service}Provider` — `gemini_provider.rb` holds
+   `GeminiProvider`. Subclass `BaseProvider` (`providers/_base_provider.rb`) or
+   an existing provider. If it needs a client gem, call
+   `require_gem!(:key, __FILE__)` at the top — the key must exist in
+   `GEM_LOADERS` at the top of `_base_provider.rb` (`:anthropic`, `:openai`,
+   `:ruby_llm` today), which is also where the gem's version requirement
+   lives, so a new client gem means a new entry there
+2. Put the supporting pieces in `lib/active_agent/providers/{service}/` — the
+   shipped providers keep `options.rb`, `request.rb`, `_types.rb` and any
+   transforms there rather than in one file (see `providers/anthropic/`)
+3. There is no registry file to edit. `provider_load`
+   (`lib/active_agent/concerns/provider.rb`) requires
+   `active_agent/providers/#{service_name.underscore}_provider` and then
+   const_gets `ActiveAgent::Providers::#{Service}Provider`, so the file name is
+   the registration. `service_name` is the config's `service:` value, or the
+   provider key camelized when the config omits it (`:openai` → `"Openai"`).
+   `PROVIDER_SERVICE_NAMES_REMAPS`, in that same file, fixes only the constant
+   half when that string doesn't camelize to your class name (`"Openai"` →
+   `"OpenAI"`); the require still uses the un-remapped name, which is why
+   `openai_provider.rb` exists next to `open_ai_provider.rb` as a one-line
+   `require_relative`. A remap without that alias file is a LoadError
 
 ## File Naming Conventions
 
@@ -186,16 +209,22 @@ end
 ## Testing
 
 ```bash
-# Run all tests
+# Run the framework's tests (bare bin/test collects test/**/*_test.rb only)
 bin/test
 
 # Run specific test file
 bin/test test/path/to/test.rb
 
 # Run tests for a specific provider
 bin/test test/integration/open_ai/
+
+# Run both trees — the framework's test/ and the engine's actionagent/test/
+bundle exec rake test
 ```
 
+The engine's tests live in `actionagent/test/` and are outside `bin/test`'s
+default glob; the Rakefile's `test` task is what sweeps both.
+
 ### Test Fixtures
 
 - VCR cassettes in `test/fixtures/vcr_cassettes/`
@@ -230,6 +259,56 @@ bin/test test/integration/open_ai/
 - Model ID determines which provider is used automatically
 - Supports prompts, embeddings, tool calling, and streaming
 
+## The dashboard: a second gem in this repo
+
+This repo ships **two** gems. `activeagent` is the framework — agents,
+providers, generation, telemetry reporting, and no Active Record. `actionagent`
+is the dashboard: a mountable Rails engine under `actionagent/`, with its own
+gemspec, holding traces and metrics, the agent builder, runs, conversations,
+evaluations, scorecards, sandboxes, session recordings, and the
+agents-as-MCP-server endpoint.
+
+They are separate because the dashboard needs `activerecord` and `solid_agent`,
+neither of which the framework requires — and `solid_agent` depends on
+`activeagent`, so the framework could never declare that second one without a
+cycle. (Both gemspecs declare `railties`; that one is not part of the split.)
+
+| What | Where |
+|------|-------|
+| Gemspec | `actionagent/actionagent.gemspec` |
+| Engine + configuration seams | `actionagent/lib/action_agent/engine.rb`, `actionagent/lib/action_agent.rb` |
+| Routes | `actionagent/config/routes.rb` |
+| Models, controllers, jobs, services, queries, serializers, views | `actionagent/app/` |
+| React source (entry `index.jsx`) | `actionagent/frontend/` |
+| Prebuilt JS/CSS (committed) | `actionagent/app/assets/builds/` |
+| Install generator | `actionagent/lib/generators/action_agent/install_generator.rb` |
+| Engine tests | `actionagent/test/` |
+
+- Everything in the engine is namespaced `ActionAgent::` (`ActionAgent::Agent`,
+  `ActionAgent::AgentRun`, `ActionAgent::AgentExecutionService`,
+  `ActionAgent::TelemetryTrace`, …). The engine is a normal gem root, so Rails
+  finds `app/` and `config/routes.rb` without help.
+- `ActionAgent::Compatibility` keeps the pre-split names resolving with a
+  deprecation, so existing initializers and already-enqueued jobs keep
+  working: `ActiveAgent::Dashboard` → `ActionAgent`,
+  `ActiveAgent::TelemetryTrace` → `ActionAgent::TelemetryTrace`,
+  `ActiveAgent::ProcessTelemetryTracesJob` →
+  `ActionAgent::ProcessTelemetryTracesJob`. It prepends a `const_missing` onto
+  `ActiveAgent` rather than aliasing eagerly, so the old names don't load the
+  engine's models — and drag Active Record in at boot — just by existing.
+- Paths are relative to wherever the host mounts the engine (the generator
+  writes `/activeagents`): `<mount>/api/...` for the JSON API, `<mount>/mcp`
+  for the MCP endpoint, `<mount>/console/traces` for the server-rendered
+  views; every other path outside `/api` renders the React app for
+  client-side routing.
+- The frontend is built with `npm run build` inside `actionagent/frontend/` and the
+  output is committed, so host apps never run a JavaScript build. Initial
+  state reaches React through a JSON data attribute, not Inertia.
+- Host integration goes through `ActionAgent.configure` seams
+  (authentication, `current_user_resolver`, `multi_tenant`,
+  `table_name_prefix`, `execution_enabled`, sandbox backends, quotas); all
+  are optional and unset means single-user self-hosted behaviour.
+
 ## Common Gotchas
 
 1. **Generation is lazy** - Nothing happens until `generate_now` or `prompt_later`
@@ -248,6 +327,11 @@ rails generate active_agent:install
 # Generate agent
 rails generate active_agent:agent MyAgent action1 action2
 
+# Install the dashboard engine (migrations, mount, initializer).
+# Needs the actionagent gem — it does not come with activeagent.
+bundle add actionagent
+rails generate action_agent:install
+
 # Run tests
 bin/test
 
@@ -260,6 +344,12 @@ bin/rubocop
 - Ruby 3.1+
 - Rails 7.2+ / 8.0+ / 8.1+
 - Provider gems (optional): `openai`, `anthropic`, `ruby_llm`
+- `activeagent` depends on actionpack, actionview, activesupport, activemodel,
+  activejob, railties and `activeagents-telemetry` — deliberately **not**
+  activerecord
+- `actionagent` (optional — the dashboard) adds `activerecord` and
+  `solid_agent` on top of `activeagent`. Installing the framework does not
+  install it: `bundle add actionagent` before running its generator
 
 ## Links
 

CHANGELOG.md

@@ -5,6 +5,59 @@ All notable changes to this project will be documented in this file.
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
+## [2.0.0] - Unreleased
+
+### ⚠️ Breaking: the dashboard is now a separate gem
+
+The dashboard engine that shipped inside `activeagent` has moved to a new
+gem, **`actionagent`**. `activeagent` is now the framework alone — it no
+longer defines `ActiveAgent::Dashboard`, and no longer pulls Active Record
+into apps that do not use it.
+
+**If you mount the dashboard, add the new gem before upgrading:**
+
+```ruby
+gem "activeagent", "~> 2.0"
+gem "actionagent", "~> 2.0"   # required if you mount the dashboard
+```
+
+Upgrading `activeagent` alone will fail at boot with
+`NameError: uninitialized constant ActiveAgent::Dashboard`, raised by your
+own initializer or by the `mount ActiveAgent::Dashboard::Engine` line in
+`config/routes.rb`. This is a major version precisely so that a
+`~> 1.1` constraint will not pick it up on its own.
+
+With `actionagent` installed, the old constants keep resolving through
+`ActionAgent::Compatibility` with a deprecation warning:
+
+- `ActiveAgent::Dashboard` → `ActionAgent`
+- `ActiveAgent::TelemetryTrace` → `ActionAgent::TelemetryTrace`
+- `ActiveAgent::ProcessTelemetryTracesJob` → `ActionAgent::ProcessTelemetryTracesJob`
+
+That last one matters beyond tidiness: Active Job serializes the class name
+into the queue payload, so jobs enqueued before the upgrade still resolve
+after it.
+
+Other changes for mounted installs:
+
+- **The server-rendered traces console moves from `/traces` to
+  `/console/traces`.** `/traces` is now the React traces view — the same
+  data, with more of it.
+- **The mount is authenticated everywhere but development and test.** The
+  sandbox API, the session-recording capture endpoints and the template
+  endpoints previously allowed anonymous access; they no longer do. The
+  `GET /api/session_recordings/demo` endpoint is removed.
+- **`current_user_method` / `current_account_method` are superseded by
+  `current_user_resolver` / `current_account_resolver`.** The engine's
+  controllers are their own base class, so a host app's `current_user`
+  helper is not available to them.
+- **An unresolved owner now scopes to nothing rather than to everything.**
+  If you configure `user_class` or `account_class`, make sure the matching
+  resolver actually returns a record, or the dashboard will show no data.
+- Existing installs upgrading from the in-gem dashboard: re-run
+  `rails generate action_agent:install`. It detects the migrations you
+  already have and emits only what is missing.
+
 ## [1.1.0] - 2026-08-12
 
 ### Dashboard — self-hosted (enterprise) mount readiness

Gemfile

@@ -4,3 +4,6 @@ gem "debug" unless ENV["CI"] == "true"
 gem "rubocop-rails-omakase"
 
 gemspec
+
+# The dashboard engine, a sibling gem in this repo.
+gemspec path: "actionagent", name: "actionagent"

README.md

@@ -25,6 +25,10 @@ Use bundler to add activeagent to your Gemfile and install:
 bundle add activeagent
 ```
 
+That is the framework — agents, providers, generation and telemetry
+reporting. The dashboard is a second gem, `actionagent`, added separately
+when you want it; see [Dashboard & Dev Console](#dashboard--dev-console).
+
 Add the generation provider gem you want to use:
 
 ```bash
@@ -127,14 +131,19 @@ development:
     service: "RubyLLM"
 ```
 
-## Dev Console & Observability
+## Dashboard & Dev Console
 
-Active Agent includes a local dev console — traces with span waterfalls,
-token usage, and per-agent metrics — as a mountable Rails engine, so you
-can watch your agents while you build:
+The dashboard is its own gem, `actionagent`: a mountable Rails engine with
+traces and span waterfalls, token usage and per-agent metrics, plus the agent
+builder, runs, conversations, evaluations, scorecards and cost estimates — so
+you can watch and drive your agents while you build. It ships separately
+because its models are Active Record models and it runs agents through
+[solid_agent](https://github.com/activeagents/solid_agent) — neither of which
+`activeagent` depends on, so an app that only runs agents installs neither.
 
 ```bash
-rails generate active_agent:dashboard:install
+bundle add actionagent
+rails generate action_agent:install
 rails db:migrate
 ```
 
@@ -145,13 +154,14 @@ telemetry:
   local_storage: true
 ```
 
-Open `/activeagents` and every generation appears as a trace. See
+The generator mounts the engine at `/activeagents` — open it and every
+generation appears as a trace. See
 [docs/framework/dashboard.md](docs/framework/dashboard.md) for
-authentication, remote ingestion, and multi-tenant mode. For production
-monitoring — evaluations, cost estimates, retention, team workspaces —
-point telemetry at the hosted platform at
-[activeagents.ai](https://activeagents.ai); every workspace starts with a
-free low-volume trial.
+authentication, remote ingestion, and multi-tenant mode. The hosted
+platform at [activeagents.ai](https://activeagents.ai) runs this same
+engine multi-tenant, adding what a hosted product has to have — accounts,
+plans, billing, quotas and managed sandboxes; every workspace starts with
+a free low-volume trial.
 
 ## Features
 
@@ -199,12 +209,12 @@ response = prompt.generate_now
 
 - [Documentation](https://docs.activeagents.ai)
 - [Getting Started Guide](https://docs.activeagents.ai/getting_started)
-- [API Reference](https://docs.activeagents.ai/docs/framework)
-- [Examples](https://docs.activeagents.ai/docs/agents)
+- [API Reference](https://docs.activeagents.ai/framework)
+- [Examples](https://docs.activeagents.ai/agents)
 
 ## Contributing
 
-We welcome contributions! Please see our [Contributing Guide](CONTRIBUTING.md) for details.
+We welcome contributions! Please see our [Contributing Guide](CONTRIBUTING.MD) for details.
 
 ## License
 

Rakefile

@@ -4,10 +4,40 @@ require "rake/testtask"
 
 Rake::TestTask.new(:test) do |t|
   t.libs << "test"
-  t.test_files = FileList["test/**/*_test.rb"]
+  t.test_files = FileList["test/**/*_test.rb", "actionagent/test/**/*_test.rb"]
     .exclude("test/**/integration_test.rb")
     .exclude("test/dummy/tmp/**/*")
   t.verbose = true
 end
 
 task default: :test
+
+# bundler/gem_tasks only discovers the gemspec at the repository root, so the
+# dashboard gem needs its own build path. It must be built from its own
+# directory: RubyGems resolves a gemspec's file list against the working
+# directory, so `gem build actionagent/actionagent.gemspec` from here would
+# not find the files it lists.
+namespace :actionagent do
+  desc "Build the actionagent gem into pkg/"
+  task :build do
+    require "fileutils"
+    root = File.expand_path(__dir__)
+    FileUtils.mkdir_p(File.join(root, "pkg"))
+
+    Dir.chdir(File.join(root, "actionagent")) do
+      sh "gem build actionagent.gemspec"
+      version = File.read("lib/action_agent/version.rb")[/VERSION = "([^"]+)"/, 1]
+      gem_file = "actionagent-#{version}.gem"
+
+      # A gem missing its own entry point is the failure mode this guards
+      # against — it installs and resolves, then dies on require.
+      contents = `tar -xOf #{gem_file} data.tar.gz | tar -tzf -`
+      %w[lib/action_agent.rb config/routes.rb app/assets/builds/action_agent.js].each do |required|
+        raise "actionagent gem is missing #{required}" unless contents.include?(required)
+      end
+
+      FileUtils.mv(gem_file, File.join(root, "pkg", gem_file))
+      puts "actionagent #{version} -> pkg/#{gem_file}"
+    end
+  end
+end

actionagent/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2025 Active Agents AI
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.