Featured Guides
The most important starting points for new and experienced users.
60-Second Quick Start
Complete the Welcome Journey, configure your workspace, and start chatting with ARMES GPT.
Auto vs. Eco Intelligence Routing
How ARMES selects frontier models or high-efficiency MoE models for maximum intelligence per watt.
Persistent Files vs Attachments
Upload documents once so every agent can summarize, search, and chunk-read them indefinitely.
The Glass-Box Memory Ledger
Inspect exactly what facts agents retain, review proposed memories, and delete anytime.
People & Messages with AI Agents
Connect privately with people you trust and invite your AI team into encrypted conversations.
Zero Data Retention Architecture
How account-level OpenRouter ZDR guarantees your prompts and completions are never retained.
AI Models & Labs Directory
In-depth technical dossiers, architecture profiles, and empirical evaluations across 21+ AI laboratories and 108+ foundation models unified under ARMES Zero Data Retention.
Architecture type, dense vs. sparse MoE, recurrent reasoning depth, and context scale up to 2M+ tokens.
Verified performance scores across MMLU-Pro, GPQA Diamond, SWE-bench Verified coding, and autonomous terminal execution.
Task-matching analysis detailing the specific coding, reasoning, drafting, and agentic workflows best suited for each engine.
Documented boundaries, latency characteristics, reasoning token overhead, and practical constraints before deploying to production.
ChatGPT
Claude
Gemini
Grok
Perplexity
DeepSeek
Mistral
MoonshotThe ARMES Knowledge Library
Organized into 8 dedicated tracks covering every surface of the platform.
Getting Started
Core philosophy, fast onboarding, multi-platform setup, and choosing the right plan.
Core Workspace
Master the Command Center, Notes knowledge base, persistent files, and project lenses.
AI Models & Intelligence
95+ models, Auto frontier routing, Eco MoE energy efficiency, and manual model pinning.
Agents & Teammates
ARMES GPT flagship, 40+ specialized agents, custom instructions, and multi-agent consultation.
Built-In Tools
Real-time Google search, cited research, unified ZDR image generation, and neural voice.
Memory & Personalization
Glass-box memory ledger, inspectable envelopes, episode summaries, and user-owned control.
People & Collaboration
Private encrypted messaging, mutual connections, invisible identity, and AI in shared rooms.
Privacy, Security & Dev
Zero Data Retention architecture, data deletion, transparent rate limits, and MCP integration.
iOS & Mobile Experience
Native iOS app, Siri voice queries, Control Center quick prompts, Lock Screen widgets, and Spotlight search.
Zero Data Retention
By Architecture, Not A Toggle.
Every AI inference request routes through OpenRouter with four-scope account-level ZDR. AI labs process prompts in memory and purge immediately. Your chats, notes, and memory remain exclusively in your personal vault until you delete them.
Zero retention across OpenAI, Anthropic, Google, and non-frontier endpoints. No training on your conversations, no human review queues, no ad profiling.
User-controlled storage in Supabase with RLS isolation and server-side encryption. What you delete is gone from active production databases.
Inspect the exact envelope your agents receive. View memory facts with provenance, keep or discard, and audit connected tools.
Connect your workspace directly to Cursor, VS Code, or Claude Code via npx armes-mcp with the same zero data retention guarantees.