<p>Magi develops a suite of AI assistant tools that serve as both practical products and testbeds for neuro-symbolic AI integration. The tools share a common MCP (Model Context Protocol) architecture, enabling AI agents to consume structured data from multiple sources.</p> <h4>Magi Assistant GM</h4> <p>An AI production stage manager for live tabletop RPG sessions, using Claude as the reasoning engine. The system consumes session data from three MCP servers (Discord bot, Foundry VTT bridge, Magi Archive wiki) and delivers real-time advice to the GM as whispered Foundry chat messages. Currently at v7 with 7 design iterations driven by live session post-mortems.</p> <p>Key capabilities: scene look-ahead from session planning cards, phonetic matching for speech-to-text garbling (Soundex + Metaphone), GM hesitation detection, anti-echo policy, flowing-RP suppression, and proactive content delivery from beat cards. Deployed to production with 92 tests across 3 files.</p> <h4>Magi Assistant Discord</h4> <p>A Discord bot providing per-user audio capture, text channel monitoring, and real-time speech-to-text (Google Cloud Speech V2). Designed to produce clean, annotated datasets for training an AI Game Master. TypeScript/Node.js with SQLite storage, MCP server for AI agent access. Phase 1 (recording) deployed; Phase 2 (transcription + MCP) code-complete.</p> <h4>Magi Assistant Foundry</h4> <p>A Foundry VTT bridge sidecar that captures Fate Core game state (actors, chat, combat, scenes) from the GM's browser via WebSocket and exposes it through an MCP server. Supports scene monitoring, status effect integration, intercut checkpoints, and experimental token proximity reporting. Architecture: browser-side Foundry module → WebSocket → Node.js sidecar → MCP.</p> <h4>Magi Assistant Smart Glasses</h4> <p>A voice-controlled wearable remote interface for AI coding agents (Claude Code, Gemini, Codex) running on VITURE Beast XR glasses. Not a standalone LLM client — it is a hands-free remote to existing development terminals, with session switching, push-to-talk, and paged terminal display. Multi-repo architecture: coordinator (Node.js), Android app (Kotlin), shared protocol library (TypeScript with codegen). Connected via Tailscale VPN.</p> <h4>Magi Archive MCP</h4> <p>MCP server for the Magi Archive wiki (Decko platform at <a href="https://wiki.magi-agi.org/">wiki.magi-agi.org</a>). Provides AI agents with search, read, create, and update access to the wiki's card-based knowledge graph. Used by the GM Assistant for runtime wiki lookups and by Claude conversations for wiki maintenance.</p>
Magi develops a suite of AI assistant tools that serve as both practical products and testbeds for neuro-symbolic AI integration. The tools share a common MCP (Model Context Protocol) architecture, enabling AI agents to consume structured data from multiple sources.
An AI production stage manager for live tabletop RPG sessions, using Claude as the reasoning engine. The system consumes session data from three MCP servers (Discord bot, Foundry VTT bridge, Magi Archive wiki) and delivers real-time advice to the GM as whispered Foundry chat messages. Currently at v7 with 7 design iterations driven by live session post-mortems.
Key capabilities: scene look-ahead from session planning cards, phonetic matching for speech-to-text garbling (Soundex + Metaphone), GM hesitation detection, anti-echo policy, flowing-RP suppression, and proactive content delivery from beat cards. Deployed to production with 92 tests across 3 files.
A Discord bot providing per-user audio capture, text channel monitoring, and real-time speech-to-text (Google Cloud Speech V2). Designed to produce clean, annotated datasets for training an AI Game Master. TypeScript/Node.js with SQLite storage, MCP server for AI agent access. Phase 1 (recording) deployed; Phase 2 (transcription + MCP) code-complete.
A Foundry VTT bridge sidecar that captures Fate Core game state (actors, chat, combat, scenes) from the GM's browser via WebSocket and exposes it through an MCP server. Supports scene monitoring, status effect integration, intercut checkpoints, and experimental token proximity reporting. Architecture: browser-side Foundry module → WebSocket → Node.js sidecar → MCP.
A voice-controlled wearable remote interface for AI coding agents (Claude Code, Gemini, Codex) running on VITURE Beast XR glasses. Not a standalone LLM client — it is a hands-free remote to existing development terminals, with session switching, push-to-talk, and paged terminal display. Multi-repo architecture: coordinator (Node.js), Android app (Kotlin), shared protocol library (TypeScript with codegen). Connected via Tailscale VPN.
MCP server for the Magi Archive wiki (Decko platform at wiki.magi-agi.org). Provides AI agents with search, read, create, and update access to the wiki's card-based knowledge graph. Used by the GM Assistant for runtime wiki lookups and by Claude conversations for wiki maintenance.
<p>Magi develops a suite of AI assistant tools that serve as both practical products and testbeds for neuro-symbolic AI integration. The tools share a common MCP (Model Context Protocol) architecture, enabling AI agents to consume structured data from multiple sources.</p> <h4>Magi Assistant GM</h4> <p>An AI production stage manager for live tabletop RPG sessions, using Claude as the reasoning engine. The system consumes session data from three MCP servers (Discord bot, Foundry VTT bridge, Magi Archive wiki) and delivers real-time advice to the GM as whispered Foundry chat messages. Currently at v7 with 7 design iterations driven by live session post-mortems.</p> <p>Key capabilities: scene look-ahead from session planning cards, phonetic matching for speech-to-text garbling (Soundex + Metaphone), GM hesitation detection, anti-echo policy, flowing-RP suppression, and proactive content delivery from beat cards. Deployed to production with 92 tests across 3 files.</p> <h4>Magi Assistant Discord</h4> <p>A Discord bot providing per-user audio capture, text channel monitoring, and real-time speech-to-text (Google Cloud Speech V2). Designed to produce clean, annotated datasets for training an AI Game Master. TypeScript/Node.js with SQLite storage, MCP server for AI agent access. Phase 1 (recording) deployed; Phase 2 (transcription + MCP) code-complete.</p> <h4>Magi Assistant Foundry</h4> <p>A Foundry VTT bridge sidecar that captures Fate Core game state (actors, chat, combat, scenes) from the GM's browser via WebSocket and exposes it through an MCP server. Supports scene monitoring, status effect integration, intercut checkpoints, and experimental token proximity reporting. Architecture: browser-side Foundry module → WebSocket → Node.js sidecar → MCP.</p> <h4>Magi Assistant Smart Glasses</h4> <p>A voice-controlled wearable remote interface for AI coding agents (Claude Code, Gemini, Codex) running on VITURE Beast XR glasses. Not a standalone LLM client — it is a hands-free remote to existing development terminals, with session switching, push-to-talk, and paged terminal display. Multi-repo architecture: coordinator (Node.js), Android app (Kotlin), shared protocol library (TypeScript with codegen). Connected via Tailscale VPN.</p> <h4>Magi Archive MCP</h4> <p>MCP server for the Magi Archive wiki (Decko platform at <a href="https://wiki.magi-agi.org/">wiki.magi-agi.org</a>). Provides AI agents with search, read, create, and update access to the wiki's card-based knowledge graph. Used by the GM Assistant for runtime wiki lookups and by Claude conversations for wiki maintenance.</p>
Magi develops a suite of AI assistant tools that serve as both practical products and testbeds for neuro-symbolic AI integration. The tools share a common MCP (Model Context Protocol) architecture, enabling AI agents to consume structured data from multiple sources.
An AI production stage manager for live tabletop RPG sessions, using Claude as the reasoning engine. The system consumes session data from three MCP servers (Discord bot, Foundry VTT bridge, Magi Archive wiki) and delivers real-time advice to the GM as whispered Foundry chat messages. Currently at v7 with 7 design iterations driven by live session post-mortems.
Key capabilities: scene look-ahead from session planning cards, phonetic matching for speech-to-text garbling (Soundex + Metaphone), GM hesitation detection, anti-echo policy, flowing-RP suppression, and proactive content delivery from beat cards. Deployed to production with 92 tests across 3 files.
A Discord bot providing per-user audio capture, text channel monitoring, and real-time speech-to-text (Google Cloud Speech V2). Designed to produce clean, annotated datasets for training an AI Game Master. TypeScript/Node.js with SQLite storage, MCP server for AI agent access. Phase 1 (recording) deployed; Phase 2 (transcription + MCP) code-complete.
A Foundry VTT bridge sidecar that captures Fate Core game state (actors, chat, combat, scenes) from the GM's browser via WebSocket and exposes it through an MCP server. Supports scene monitoring, status effect integration, intercut checkpoints, and experimental token proximity reporting. Architecture: browser-side Foundry module → WebSocket → Node.js sidecar → MCP.
A voice-controlled wearable remote interface for AI coding agents (Claude Code, Gemini, Codex) running on VITURE Beast XR glasses. Not a standalone LLM client — it is a hands-free remote to existing development terminals, with session switching, push-to-talk, and paged terminal display. Multi-repo architecture: coordinator (Node.js), Android app (Kotlin), shared protocol library (TypeScript with codegen). Connected via Tailscale VPN.
MCP server for the Magi Archive wiki (Decko platform at wiki.magi-agi.org). Provides AI agents with search, read, create, and update access to the wiki's card-based knowledge graph. Used by the GM Assistant for runtime wiki lookups and by Claude conversations for wiki maintenance.
<p><strong>Sources consulted:</strong> Magi Archive <code>Tools+Magi Assistant GM</code> (ID 11072, updated 2026-03-21); Magi Archive <code>Tools+Magi Assistant GM+Architecture</code> (ID 11074, updated 2026-03-21); Hyperon RawData <code>RawData+transcripts+Magi Weekly+2026-02-16</code>; current Hyperon Wiki parent card <code>Ecosystem+Magi+Magi Deep Dive+Tools and Assistants</code> (ID 7184, RichText, updated 2026-06-05); Magi Archive <code>The Plan+Wiki Finalization Meta-Plan</code> WS6 section (ID 17982, updated 2026-06-27).</p> <h3>Summary of proposed correction</h3> <p>The parent card currently presents Magi assistant tools as practical products and testbeds for neuro-symbolic AI integration. Source evidence confirms the GM Assistant and related current tools are real LLM/MCP runtime tools — notably a TypeScript + Node.js service using Claude, Discord, Foundry VTT, and Magi Archive wiki MCP connections — but does not show current Hyperon, MeTTa, or MAGUS reasoning integration in those assistant tools. The proposed correction keeps the implementation status of the assistant tooling while removing the unsupported implication that the current suite is already a Hyperon/MAGUS neuro-symbolic integration testbed.</p> <h3>Proposal P-1 — Replace the opening assistant-suite framing</h3> <p><strong>Target card:</strong> <code>Ecosystem+Magi+Magi Deep Dive+Tools and Assistants</code></p> <p><strong>Current content:</strong></p> <blockquote><p>Magi develops a suite of AI assistant tools that serve as both practical products and testbeds for neuro-symbolic AI integration. The tools share a common MCP (Model Context Protocol) architecture, enabling AI agents to consume structured data from multiple sources.</p></blockquote> <p><strong>Proposed replacement:</strong></p> <blockquote><p>Magi develops a suite of AI assistant tools that serve as practical LLM/MCP runtime integrations for gameplay, coding, and wiki workflows. These tools share a common Model Context Protocol architecture, enabling AI agents to consume structured data from Discord, Foundry VTT, Magi Archive, and related sources. Current evidence supports treating them as operational assistant/tooling infrastructure; Hyperon, MeTTa, and MAGUS reasoning integration should be described as future or separate work unless source-specific integration evidence is added.</p></blockquote> <p><strong>Rationale:</strong> Magi Archive <code>Tools+Magi Assistant GM</code> describes the GM Assistant as Claude-based production-stage-manager tooling consuming Discord, Foundry bridge, and Magi Archive wiki data. <code>Tools+Magi Assistant GM+Architecture</code> describes a TypeScript + Node.js service with three MCP servers and Claude as backend; exact searches in that architecture card found no <code>Hyperon</code>, and exact searches in the main GM card found no <code>Hyperon</code> or <code>MAGUS</code>. The 2026-02-16 Magi Weekly transcript confirms live gameplay integration through Discord and Foundry, but likewise supports LLM/MCP assistant tooling rather than Hyperon/MAGUS reasoning infrastructure.</p> <h3>Proposal P-2 — Keep GM Assistant status, but avoid over-reading it as MAGUS/Hyperon integration</h3> <p><strong>Target paragraph:</strong> <code>Magi Assistant GM</code> section.</p> <p><strong>Current content:</strong></p> <blockquote><p>An AI production stage manager for live tabletop RPG sessions, using Claude as the reasoning engine. The system consumes session data from three MCP servers (Discord bot, Foundry VTT bridge, Magi Archive wiki) and delivers real-time advice to the GM as whispered Foundry chat messages. Currently at v7 with 7 design iterations driven by live session post-mortems.</p></blockquote> <p><strong>Proposed replacement:</strong></p> <blockquote><p>An AI production stage manager for live tabletop RPG sessions, using Claude as the current reasoning backend. The system consumes session data from three MCP servers (Discord bot, Foundry VTT bridge, Magi Archive wiki) and delivers real-time advice to the GM as whispered Foundry chat messages. Current Magi Archive architecture notes describe this as a TypeScript + Node.js LLM/MCP service, with seven design iterations driven by live-session post-mortems.</p></blockquote> <p><strong>Rationale:</strong> This preserves the real implementation and live-session evidence while making clear that the current reasoning backend is Claude/LLM-based rather than Hyperon/MAGUS-based.</p> <h3>WS6 merge note</h3> <p>This is intended as a WS6 proposal child. If approved, a human reviewer should reconcile it against the current parent content in the merge workbench. The parent should be written only through the WS6 verifying apply gate, not by direct overwrite.</p>
Sources consulted: Magi Archive Tools+Magi Assistant GM (ID 11072, updated 2026-03-21); Magi Archive Tools+Magi Assistant GM+Architecture (ID 11074, updated 2026-03-21); Hyperon RawData RawData+transcripts+Magi Weekly+2026-02-16; current Hyperon Wiki parent card Ecosystem+Magi+Magi Deep Dive+Tools and Assistants (ID 7184, RichText, updated 2026-06-05); Magi Archive The Plan+Wiki Finalization Meta-Plan WS6 section (ID 17982, updated 2026-06-27).
The parent card currently presents Magi assistant tools as practical products and testbeds for neuro-symbolic AI integration. Source evidence confirms the GM Assistant and related current tools are real LLM/MCP runtime tools — notably a TypeScript + Node.js service using Claude, Discord, Foundry VTT, and Magi Archive wiki MCP connections — but does not show current Hyperon, MeTTa, or MAGUS reasoning integration in those assistant tools. The proposed correction keeps the implementation status of the assistant tooling while removing the unsupported implication that the current suite is already a Hyperon/MAGUS neuro-symbolic integration testbed.
Target card: Ecosystem+Magi+Magi Deep Dive+Tools and Assistants
Current content:
Magi develops a suite of AI assistant tools that serve as both practical products and testbeds for neuro-symbolic AI integration. The tools share a common MCP (Model Context Protocol) architecture, enabling AI agents to consume structured data from multiple sources.
Proposed replacement:
Magi develops a suite of AI assistant tools that serve as practical LLM/MCP runtime integrations for gameplay, coding, and wiki workflows. These tools share a common Model Context Protocol architecture, enabling AI agents to consume structured data from Discord, Foundry VTT, Magi Archive, and related sources. Current evidence supports treating them as operational assistant/tooling infrastructure; Hyperon, MeTTa, and MAGUS reasoning integration should be described as future or separate work unless source-specific integration evidence is added.
Rationale: Magi Archive Tools+Magi Assistant GM describes the GM Assistant as Claude-based production-stage-manager tooling consuming Discord, Foundry bridge, and Magi Archive wiki data. Tools+Magi Assistant GM+Architecture describes a TypeScript + Node.js service with three MCP servers and Claude as backend; exact searches in that architecture card found no Hyperon, and exact searches in the main GM card found no Hyperon or MAGUS. The 2026-02-16 Magi Weekly transcript confirms live gameplay integration through Discord and Foundry, but likewise supports LLM/MCP assistant tooling rather than Hyperon/MAGUS reasoning infrastructure.
Target paragraph: Magi Assistant GM section.
Current content:
An AI production stage manager for live tabletop RPG sessions, using Claude as the reasoning engine. The system consumes session data from three MCP servers (Discord bot, Foundry VTT bridge, Magi Archive wiki) and delivers real-time advice to the GM as whispered Foundry chat messages. Currently at v7 with 7 design iterations driven by live session post-mortems.
Proposed replacement:
An AI production stage manager for live tabletop RPG sessions, using Claude as the current reasoning backend. The system consumes session data from three MCP servers (Discord bot, Foundry VTT bridge, Magi Archive wiki) and delivers real-time advice to the GM as whispered Foundry chat messages. Current Magi Archive architecture notes describe this as a TypeScript + Node.js LLM/MCP service, with seven design iterations driven by live-session post-mortems.
Rationale: This preserves the real implementation and live-session evidence while making clear that the current reasoning backend is Claude/LLM-based rather than Hyperon/MAGUS-based.
This is intended as a WS6 proposal child. If approved, a human reviewer should reconcile it against the current parent content in the merge workbench. The parent should be written only through the WS6 verifying apply gate, not by direct overwrite.