@rizom/brain 0.2.0-alpha.6 → 0.2.0-alpha.60

package/dist/seed-content/relay/README.md

@@ -1,384 +1,41 @@
-# Recall
+# Relay seed content
 
-**Your team's AI-powered knowledge hub**
+This directory contains the first-boot starter content for `@brains/relay`.
 
-Recall captures links, conversations, and decisions automatically, then makes them searchable and actionable. Connect via Matrix chat, web interface, or AI assistants like Claude Desktop.
+## What it is for
 
-## What is Recall?
+The files here give a new relay instance enough durable markdown content to boot with a coherent identity and a usable starting structure.
 
-Recall is a collaborative knowledge management system that:
+Included content typically covers:
 
-- **Captures** links, conversations, and documents automatically
-- **Organizes** content with AI-powered summaries and topic extraction
-- **Connects** to Matrix chat, web browsers, and AI assistants
-- **Syncs** everything to Git for version control and backup
-- **Stores** data as human-readable Markdown files
+- `brain-character/` — the brain's identity and operating voice
+- `anchor-profile/` — the operator or owner profile
+- `site-info/` — site title, metadata, and presentation defaults
+- `site-content/` — editable homepage and about-page sections for the default site
+- `deck/` — example deck content
+- root `.md` files — supporting reference material
 
-Think of it as a smart second brain for your team that works where you already collaborate.
+## How it is used
 
-## Features
+On first boot, if `brain-data/` is empty, relay copies this seed content into the instance content directory.
 
-### 🔗 Automatic Link Capture
+After that:
 
-Paste any URL in Matrix chat and Recall automatically:
+- edit the instance's `brain-data/` content, not this directory
+- treat markdown files as the durable source of truth
+- let the runtime index that content into SQLite for search and retrieval
+- optionally sync the same markdown content to git through `directory-sync`
 
-- Extracts and archives the content
-- Generates a summary
-- Identifies relevant topics
-- Makes it searchable
+## Authoring notes
 
-### 💬 AI-Powered Conversations
+- Keep content markdown-first and entity-shaped
+- Prefer durable editorial copy over temporary scaffolding text
+- Use frontmatter that matches the owning entity schema
+- Keep examples realistic but clearly starter-level
 
-- Ask questions in natural language via Matrix
-- Query your team's knowledge base
-- Get instant answers from captured content
-- Integrated semantic search
+## Related docs
 
-### 🏷️ Smart Organization
-
-- Automatic topic extraction from conversations
-- AI-generated summaries of discussions
-- Browse by links, topics, summaries, or presentations
-- Full-text and semantic search
-
-### 🌐 Static Site Generation
-
-- Beautiful web interface with all your content
-- Presentation decks with Reveal.js
-- Responsive design with Tailwind CSS
-- Can be deployed anywhere
-
-### 🔄 Git Sync
-
-- All content backed up to Git automatically
-- Edit in Obsidian, VS Code, or any Markdown editor
-- Version control for all changes
-- Works with GitHub, GitLab, or self-hosted Git
-
-### 🤖 MCP Integration
-
-- Connect to Claude Desktop or other MCP clients
-- Let AI assistants search your knowledge base
-- Execute commands and queries remotely
-- HTTP and stdio transport support
-
-## Quick Start
-
-### Prerequisites
-
-- [Bun](https://bun.sh) v1.0+
-- SQLite 3.x
-- Git
-
-### Installation
-
-```bash
-# Clone the repository
-git clone https://github.com/your-org/brains
-cd brains
-
-# Install dependencies
-bun install
-
-# Navigate to your relay app
-cd apps/my-relay
-```
-
-### Configuration
-
-Create a `.env` file in the root directory:
-
-```bash
-# Required: Anthropic API key for AI features
-AI_API_KEY=your-api-key-here
-
-# Database URLs (defaults to local SQLite files)
-DATABASE_URL=file:./data/brains.db
-JOB_QUEUE_DATABASE_URL=file:./data/jobs.db
-CONVERSATION_DATABASE_URL=file:./data/conversations.db
-
-# Matrix Configuration (optional - for chat interface)
-MATRIX_HOMESERVER=https://matrix.org
-MATRIX_ACCESS_TOKEN=your_matrix_token
-MATRIX_USER_ID=@recall:matrix.org
-MATRIX_ANCHOR_USER_ID=@youruser:matrix.org
-
-# Git Sync Configuration (optional - for backup)
-GIT_SYNC_REPO=your-org/recall-content
-GIT_SYNC_TOKEN=ghp_your_github_token
-
-# MCP Configuration (optional - for remote access)
-MCP_AUTH_TOKEN=your_secure_token
-
-# Web Interface (optional - for production deployment)
-DOMAIN=recall.example.com
-PORT=3000
-
-# Logging
-LOG_LEVEL=info
-```
-
-### Running Locally
-
-```bash
-# Run in development mode (watches for changes)
-bun run dev
-
-# Run CLI interface
-bun run dev:cli
-
-# Run in production mode
-bun run start
-```
-
-The application will:
-
-- Initialize SQLite databases
-- Sync seed content to markdown files
-- Start all configured interfaces (web, matrix, MCP)
-- Begin processing jobs in the background
-
-Visit `http://localhost:3000` to access the web interface.
-
-## Development
-
-### Available Commands
-
-```bash
-# Development
-bun run dev              # Start with hot reload
-bun run dev:cli          # Start CLI interface with hot reload
-
-# Production
-bun run start            # Run the application
-bun run start:cli        # Run CLI interface
-
-# Database
-bun run migrate          # Run all migrations
-bun run migrate:entities # Migrate entities database only
-bun run db:studio        # Open Drizzle Studio (main DB)
-bun run db:studio:entities    # Studio for entities
-bun run db:studio:conversations # Studio for conversations
-bun run db:studio:jobs   # Studio for job queue
-
-# Quality
-bun run test             # Run tests
-bun run typecheck        # Type check
-bun run lint             # Lint code
-
-# Matrix Setup
-bun run matrix:setup     # Interactive Matrix configuration
-```
-
-### Project Structure
-
-```
-apps/my-relay/
-├── brain.config.ts      # Main configuration
-├── seed-content/        # Initial markdown content
-├── scripts/             # Utility scripts
-└── data/                # SQLite databases (created on first run)
-```
-
-### Adding Seed Content
-
-Place markdown files in `seed-content/` and they'll be synced to the database on startup. Files can include:
-
-- Links (frontmatter: `type: link`)
-- Summaries (frontmatter: `type: summary`)
-- Topics (frontmatter: `type: topic`)
-- Presentations (frontmatter: `type: deck`)
-- Custom content types
-
-## Usage
-
-### Matrix Chat
-
-1. Invite your Recall bot to a Matrix room
-2. Paste URLs to capture links automatically
-3. Mention the bot to ask questions: `@recall what topics have we discussed?`
-4. Use commands: `!help`, `!identity`, `!git-sync`
-
-### Web Interface
-
-Navigate to your deployment URL to:
-
-- Browse all captured links at `/links`
-- View topics at `/topics`
-- Read summaries at `/summaries`
-- Present decks at `/decks`
-- Search content with semantic search
-
-### MCP (AI Assistant Integration)
-
-Configure Claude Desktop or another MCP client:
-
-```json
-{
-  "mcpServers": {
-    "recall": {
-      "command": "node",
-      "args": ["/path/to/brains/apps/my-relay/brain.yaml", "--mcp"]
-    }
-  }
-}
-```
-
-Or use HTTP transport for remote access:
-
-```json
-{
-  "mcpServers": {
-    "recall": {
-      "url": "https://recall.example.com/mcp",
-      "headers": {
-        "Authorization": "Bearer your_auth_token"
-      }
-    }
-  }
-}
-```
-
-Available MCP tools:
-
-- `brain_query` - Natural language queries
-- `brain_command` - Execute shell commands
-- `entity_*` - Create, read, update, delete entities
-
-### CLI Interface
-
-```bash
-bun run start:cli
-
-# In the CLI:
-> query "what are our main topics?"
-> command "help"
-> exit
-```
-
-## Deployment
-
-Recall can be deployed as:
-
-1. **Docker Container** - See `deploy/providers/hetzner/` for examples
-2. **Systemd Service** - Run as a system service
-3. **Node.js Process** - Standard Node deployment
-4. **Bun Binary** - Compile to standalone executable
-
-### Docker Deployment
-
-```bash
-# Build image
-docker build -t recall .
-
-# Run container
-docker run -d \
-  --name recall \
-  -p 3000:3000 \
-  -v ./data:/app/data \
-  -v ./content:/app/content \
-  --env-file .env \
-  recall
-```
-
-### Environment-Specific Configuration
-
-- **Development**: Local SQLite, hot reload, debug logging
-- **Production**: Remote databases, health checks, info logging
-- **Preview**: Separate databases, preview domain
-
-See `deploy/` directory for deployment scripts and configurations.
-
-## Architecture
-
-Recall is built on a plugin-based architecture:
-
-### Core
-
-- **Shell** - Plugin system and service orchestration
-- **Entity Service** - Unified data model with SQLite + vector embeddings
-- **Job Queue** - Background task processing
-- **Messaging** - Event-driven communication between plugins
-
-### Plugins
-
-- **System** - Core commands and queries
-- **Link** - URL capture and content extraction
-- **Summary** - AI-powered summarization
-- **Topics** - Topic extraction and organization
-- **Decks** - Presentation deck rendering
-- **Site Builder** - Static site generation
-- **Directory Sync** - Markdown file synchronization
-- **Git Sync** - Automatic Git backup
-
-### Interfaces
-
-- **Matrix** - Chat integration
-- **MCP** - AI assistant protocol
-- **Webserver** - HTTP API and static site hosting
-- **CLI** - Command-line interface
-
-## Contributing
-
-This is an internal tool, but we welcome contributions:
-
-1. Follow the repository [AGENTS.md](../../AGENTS.md) development guidelines
-2. Write tests for new features
-3. Run `bun run typecheck` and `bun run lint` before committing
-4. Keep commits focused and write clear messages
-
-## Troubleshooting
-
-### Database Issues
-
-```bash
-# Reset databases (WARNING: deletes all data)
-rm -rf data/*.db
-
-# Run migrations
-bun run migrate
-```
-
-### Matrix Connection Issues
-
-```bash
-# Verify Matrix configuration
-bun run matrix:setup
-
-# Check Matrix access token validity
-curl -H "Authorization: Bearer $MATRIX_ACCESS_TOKEN" \
-  $MATRIX_HOMESERVER/_matrix/client/v3/account/whoami
-```
-
-### MCP Connection Issues
-
-```bash
-# Test MCP endpoint
-curl -X POST https://your-domain.com/mcp \
-  -H "Authorization: Bearer $MCP_AUTH_TOKEN" \
-  -H "Content-Type: application/json" \
-  -d '{"method": "ping"}'
-```
-
-### Git Sync Issues
-
-```bash
-# Check Git sync status
-cd content
-git status
-git log
-
-# Force sync
-# In Matrix: !git-sync
-# Or restart the application
-```
-
-## License
-
-Apache-2.0
-
-## Support
-
-- Documentation: [docs/](../../docs/)
-- Issues: [GitHub Issues](https://github.com/your-org/brains/issues)
-- Matrix: `#recall:your-domain.com`
+- [Relay model README](../README.md)
+- [Repository README](../../../README.md)
+- [Brain model architecture](../../../docs/brain-model.md)
+- [AGENTS.md](../../../AGENTS.md)

package/dist/seed-content/relay/anchor-profile/anchor-profile.md

@@ -1,17 +1,14 @@
 ---
 kind: team
-name: Team Organization
-description: Collaborative team workspace for knowledge sharing
-website: "https://team.example.com"
-email: team@example.com
+name: Relay POC Team
+description: A small team validating shared knowledge capture, synthesis, and peer-brain coordination.
+website: "https://relay.example.com"
+email: relay-team@example.com
 socialLinks:
-  - platform: linkedin
-    url: "https://www.linkedin.com/company/team-org"
-    label: Connect on LinkedIn
   - platform: github
-    url: "https://github.com/team-org"
-    label: View team projects
+    url: "https://github.com/rizom-ai/brains"
+    label: Brains repository
   - platform: email
-    url: "mailto:team@example.com"
-    label: Contact team
+    url: "mailto:relay-team@example.com"
+    label: Contact the Relay POC team
 ---

package/dist/seed-content/relay/brain-character/brain-character.md

@@ -1,9 +1,10 @@
 ---
 name: Relay
 role: Team knowledge coordinator
-purpose: Capture, organize, and share team knowledge
+purpose: Capture, synthesize, and route shared team knowledge across humans and peer brains
 values:
-  - clarity
-  - collaboration
-  - accessibility
+  - collaborative clarity
+  - durable memory
+  - grounded synthesis
+  - explicit coordination
 ---

package/dist/seed-content/relay/site-content/about/about.md

@@ -0,0 +1,20 @@
+---
+routeId: about
+sectionId: about
+---
+
+# About Relay
+
+## Title
+
+Relay is a collaborative team-memory brain.
+
+## Intro
+
+It exists for teams that need to remember together: capturing what happened, summarizing why it matters, and coordinating with trusted peer brains when work crosses boundaries.
+
+## Points
+
+- Not a personal blog engine or portfolio shell.
+- Not a marketing automation stack.
+- A small public face for a larger private knowledge workflow.

package/dist/seed-content/relay/site-content/home/diagram.md

@@ -0,0 +1,184 @@
+---
+routeId: home
+sectionId: diagram
+---
+
+# Home diagram
+
+## Eyebrow
+
+A team brain, diagrammed
+
+## Headline
+
+Relay sits between the work and the world, and keeps both honest.
+
+## Intro
+
+Capture sources flow into a shared brain; the brain organizes, summarizes, and — selectively — exposes a public surface. Everything else stays private and operational.
+
+## Primary CTA
+
+### Label
+
+See it on a real team
+
+### Href
+
+#diagram
+
+## Secondary CTA
+
+### Label
+
+Read the model
+
+### Href
+
+/about
+
+## Inputs
+
+### Input 1
+
+#### Label
+
+Source · chat
+
+#### Title
+
+Discord
+
+#### Detail
+
+Shared decisions and field notes captured in flow.
+
+### Input 2
+
+#### Label
+
+Source · agent
+
+#### Title
+
+MCP / CLI
+
+#### Detail
+
+Structured captures from tools and assistants.
+
+### Input 3
+
+#### Label
+
+Source · web
+
+#### Title
+
+Links & docs
+
+#### Detail
+
+External material indexed with sourceable metadata.
+
+## Outputs
+
+### Output 1
+
+#### Label
+
+Surface · public
+
+#### Title
+
+Default site
+
+#### Detail
+
+A small, durable face onto what the team currently knows.
+
+### Output 2
+
+#### Label
+
+Surface · agents
+
+#### Title
+
+A2A exchange
+
+#### Detail
+
+Peer brains coordinate over an approved, signed protocol.
+
+### Output 3
+
+#### Label
+
+Surface · query
+
+#### Title
+
+Team Q&A
+
+#### Detail
+
+Ask the brain in chat; answers cite the captures behind them.
+
+## Core
+
+### Eyebrow
+
+The relay
+
+### Name
+
+brain
+
+### Sub
+
+capture → topics → summaries
+
+## Legend
+
+### Legend 1
+
+#### Tone
+
+capture
+
+#### Title
+
+Capture
+
+#### Text
+
+Anything the team already does that produces a trace — chat messages, links, deploys, decisions.
+
+### Legend 2
+
+#### Tone
+
+synthesis
+
+#### Title
+
+Synthesis
+
+#### Text
+
+The work the brain does on its own time — clustering captures into topics, summaries, and durable memory.
+
+### Legend 3
+
+#### Tone
+
+share
+
+#### Title
+
+Share
+
+#### Text
+
+A small, opinionated public surface and an approved agent-to-agent protocol. Most memory stays private.

package/dist/seed-content/relay/site-info/site-info.md

@@ -1,5 +1,15 @@
 ---
 title: Relay
-description: Your team's AI-powered knowledge hub
+description: Collaborative team memory, synthesis, and peer-brain coordination.
 themeMode: dark
+copyright: Relay sample site — powered by Rizom
+cta:
+  heading: Start with shared memory
+  buttonText: Read the model
+  buttonLink: /about
+sections:
+  loop:
+    blurb: Capture, synthesize, and share what the team already knows.
+  surface:
+    blurb: Minimal by default, extensible in the full preset.
 ---