Eve library

Iris -- Product Features

Guide

Revision 1

On this page

Canonical source: the canonical features document for this domain is deep-dive/features.md, rendered in the docs center. This file is its operational companion (service topology, APIs, schemas, runtime detail) — reconciled by linking, not duplicated (§2, §13).

Iris is a Universal Intelligent Assistant Platform that provides enterprise-grade AI conversations powered by multiple language models (Claude, GPT, Gemini, local models), deep personalization through hierarchical memory, autonomous agents with tool use, knowledge base integration, multimodal interaction (text, voice, vision, spatial), privacy controls, and seamless availability across web, desktop, mobile, wearable, and XR devices. This document describes every user-facing feature from the perspective of the people who use the platform.


Table of Contents#

  1. Platform Access
  2. AI Conversations
  3. AI Model Selection
  4. Memory System
  5. Knowledge Base
  6. AI Agents
  7. Code Intelligence
  8. Voice Interaction
  9. Vision and Multimodal
  10. Emotional Intelligence
  11. Reasoning and Thinking
  12. Agent Marketplace
  13. Privacy and Security
  14. Accessibility
  15. Personalization
  16. Analytics and Insights
  17. Notifications and Presence
  18. Platform Administration
  19. Integrations
  20. Developer Platform
  21. Wearable and XR

Platform Access#

Iris is available everywhere you work and live.

Applications#

Application Platform Description
Web App Browser (React) Full-featured web interface with conversation management
Desktop App Windows/Mac/Linux (Tauri) Native desktop assistant with system integration
Mobile App iOS/Android (React Native) On-the-go AI assistance with voice and camera integration
Wearable App Smartwatches Quick AI interactions from your wrist
XR App VR/AR headsets (WebXR) Spatial AI assistant in extended reality environments
Dashboard Browser Administration and analytics dashboard
Developer Portal Browser API documentation, key management, and developer tools
Marketplace Browser Browse and install AI agents, plugins, and integrations
API Service Server REST, WebSocket, and gRPC APIs for programmatic access

Interface Overview#

Element Description
Conversation List Sidebar showing all past conversations with search
Agent Selector Choose which AI agent to use for the current conversation
Message Input Text field with file attachment and voice input buttons
Streaming Display AI responses appear in real time as they are generated
Knowledge Panel Access and manage your uploaded documents and knowledge bases
Settings Memory management, privacy controls, model preferences

Keyboard Shortcuts#

Shortcut Action
Enter Send message
Shift + Enter New line in message
Ctrl/Cmd + N New conversation
Ctrl/Cmd + K Quick search across conversations
Escape Cancel current response generation
Ctrl/Cmd + Up Edit last message
Ctrl/Cmd + / Show all keyboard shortcuts

AI Conversations#

The core of Iris is natural, context-aware AI conversation.

Conversation Management#

Feature Description
Create Conversations Start new conversations on any topic
Conversation History Complete searchable history of all past conversations
Conversation Titles Auto-generated or custom titles for easy identification
Conversation Search Full-text search across all conversations and messages
Conversation Export Export conversations as Markdown, JSON, or PDF
Conversation Deletion Delete individual conversations or bulk clear history
Conversation Branching Branch a conversation to explore a different direction

Message Features#

Feature Description
Real-Time Streaming Responses stream token by token for immediate feedback
Message Editing Edit your messages and regenerate AI responses
Response Regeneration Regenerate the AI response if you want a different answer
Message Copying One-click copy of any message or code block
File Attachments Attach documents, images, code files, and data for AI analysis
Code Block Formatting Code displayed with syntax highlighting and copy button
Markdown Rendering Rich formatting with headings, lists, tables, and links
LaTeX Support Mathematical equations rendered with LaTeX
Citation Links AI responses include citations to source documents
Stop Generation Stop AI response generation at any point

Supported File Types#

Category Formats Max Size
Documents PDF, DOCX, TXT, Markdown 25 MB
Code Python, JS, TS, Go, Rust, etc. 10 MB
Data CSV, JSON, XLSX 50 MB
Images PNG, JPG, GIF, WebP 20 MB
Slides PPTX 25 MB

AI Model Selection#

Iris provides intelligent access to multiple AI models and automatically selects the best one for each task.

Available Models#

Provider Models Strengths
Anthropic Claude (Opus, Sonnet, Haiku) Analysis, reasoning, safety
OpenAI GPT-4, GPT-4 Turbo Versatility, creative writing
Google Gemini (Pro, Ultra) Long context, multimodal
Cohere Command R+ Embeddings, search, reranking
Mistral Mistral models Fast inference, multilingual
Local Ollama (Llama, Mistral, Phi, etc.) Privacy, offline, zero-cost

Intelligent Model Routing#

Feature Description
Task-Based Routing Automatically select the best model for each type of task
Cost Optimization Route to the most cost-effective model that meets quality needs
Latency Optimization Choose faster models when response speed is critical
Fallback Chains Automatically try alternative models if the primary is unavailable
Circuit Breaking Temporarily stop routing to models experiencing errors
Provider Health Monitoring Continuous monitoring of each model provider's availability
Manual Override Users can manually select a specific model when desired

Multi-Model Ensemble#

Feature Description
Ensemble Voting Query multiple models and use voting to determine the best answer
Response Aggregation Combine insights from multiple models into a single response
Confidence Scoring Each response includes a confidence score from the model
Self-Consistency Checks Verify response consistency across multiple model runs

Memory System#

Iris remembers you across conversations through a four-tier memory system.

Short-Term Memory#

Feature Description
Current Conversation Full context of the active conversation
Recent Messages The most recent exchanges for immediate context
Redis-Backed Fast in-memory storage with 30-minute TTL
Automatic Cleanup Old short-term memories expire automatically

Working Memory#

Feature Description
Active Task Context The AI maintains context about the current task you are working on
File References Tracks files mentioned or uploaded during the session
Temporary State Session-scoped state that persists until you end the session

Long-Term Memory#

Feature Description
User Preferences Remembers your communication style, coding preferences, etc.
Learned Facts Facts you tell the AI are remembered permanently
Explicit Teaching Say "Remember that I prefer TypeScript" and Iris stores it
Confidence Scoring Each memory has a confidence score that improves over time
Source Attribution Each memory links back to the conversation that created it
Manual Memory Management View, edit, and delete stored memories in Settings

Episodic Memory#

Feature Description
Conversation Summaries Key conversations are summarized and stored
Key Moments Important interactions and breakthroughs are flagged
Milestone Tracking Track significant accomplishments across conversations
90-Day Retention Episodic memories retain for 90 days by default

Semantic Memory#

Feature Description
User Knowledge Graph Graph of entities and relationships specific to you
Domain Expertise Model Iris learns your expertise level in different domains
Vector Embeddings Semantic memories stored as vectors in Qdrant for similarity search
Memory Decay Unused memories gradually decrease in relevance

Memory Operations#

Feature Description
Memory Consolidation Short-term memories are consolidated into long-term storage
Memory Visualization Visual graph showing your memory network
Memory Debugging Tools to inspect and troubleshoot memory behavior
Memory Migration Import/export memories when switching accounts or instances
Memory Sharing Selectively share memory contexts (e.g., team knowledge)
Memory Analytics Insights into how memory improves your AI experience
Privacy Controls Granular control over what is remembered and stored

Knowledge Base#

Upload your documents and let Iris search them for informed answers.

Document Management#

Feature Description
Document Upload Upload PDFs, Word docs, Markdown, code files, and more
Bulk Upload Upload multiple documents simultaneously
Collection Organization Group related documents into named collections
Tagging Add tags to documents for filtering and organization
Version Control Track document versions with automatic change detection

Knowledge Features#

Feature Description
Semantic Search Meaning-based search that finds relevant passages by concept
Hybrid Retrieval Combines semantic search with keyword matching for best results
Automatic Chunking Documents are intelligently split into searchable chunks
Citation Engine AI responses cite specific documents, pages, and sections
Knowledge Graph Entities and relationships extracted and linked across documents
Real-Time Knowledge Access to current information through real-time data sources
Knowledge Freshness Automatic detection of outdated documents and content
Fact Checking AI cross-references answers against your knowledge base
Knowledge Curation Tools for curating and maintaining knowledge base quality
Knowledge Export Export your knowledge base in standard formats

Enterprise Knowledge#

Feature Description
Team Knowledge Bases Shared knowledge bases for teams and organizations
Access Control Fine-grained permissions on who can view and edit knowledge
Knowledge Analytics Track which documents are most referenced and useful
Grounding Ensure AI responses are grounded in your knowledge base
Personal Knowledge Base Individual knowledge bases that are private to each user

AI Agents#

Specialized AI agents designed for specific tasks and workflows.

Built-In Agents#

Agent Specialty
General Everyday questions, writing, brainstorming, learning
CodeAssist Programming, debugging, code review, architecture
Researcher Deep research, source synthesis, fact verification
Writer Content creation, editing, copywriting, style adaptation
Analyst Data analysis, report generation, insights extraction
Support Customer support, FAQ responses, ticket handling
Creative Creative writing, ideation, artistic concepts
Operations DevOps, infrastructure, monitoring, automation
Supervisor Oversees multi-agent collaboration and quality control

Agent Capabilities#

Capability Description
Task Decomposition Complex tasks are broken into manageable sub-tasks
Plan Generation Agents create step-by-step plans before executing
Tool Use Agents can use tools (file operations, web search, code execution)
Multi-Agent Collaboration Multiple agents work together on complex tasks
Proactive Suggestions Agents proactively suggest actions based on context
Workflow Automation Define repeatable workflows that agents execute automatically

Agent Tools#

Tool Category Tools Available
File Ops Read, write, search, and organize files
Web Ops Browse websites, search the internet, extract data
Code Exec Run code, debug, test in sandboxed environments
API Ops Make REST, GraphQL, and gRPC calls
Database Ops Query databases, inspect schemas, run migrations
GUI Ops Click, type, take screenshots for computer use

Agent-to-Agent Protocol (A2A)#

Feature Description
Agent Registry Discover and connect with available agents
Agent Communication Agents send messages and share context with each other
Task Delegation Agents delegate sub-tasks to specialized agents
Result Aggregation Supervisor agent aggregates results from multiple agents

Model Context Protocol (MCP)#

Feature Description
MCP Server Iris acts as an MCP server for tool and context sharing
MCP Client Iris can connect to external MCP servers for additional tools
Standard Protocol Compatible with the emerging MCP standard ecosystem

Code Intelligence#

Specialized capabilities for developers and engineering teams.

Code Features#

Feature Description
Code Generation Generate code from natural language descriptions
Code Explanation Explain what existing code does in plain language
Code Review Automated code review for bugs, security, and best practices
Code Debugging Identify and fix bugs with suggested corrections
Code Refactoring Suggest and apply code refactoring improvements
Test Generation Generate unit tests for existing code
Documentation Generation Create documentation from code
Architecture Advice Guidance on system design and architectural decisions

Codebase Understanding#

Feature Description
Repository Analysis Analyze an entire repository to understand its structure
Dependency Mapping Map out dependencies and their relationships
Code Search Semantic search across codebases for relevant code
Static Analysis Identify potential issues without running the code

Voice Interaction#

Speak to Iris naturally for hands-free AI assistance.

Voice Features#

Feature Description
Voice Input Speak your messages instead of typing
Voice Response Hear AI responses read aloud with natural voice synthesis
Continuous Conversation Speak back and forth in a natural voice conversation
Wake Word Detection Activate Iris with a custom wake word (desktop/mobile)
Multi-Language Voice Voice input and output in multiple languages
Voice Commands Execute actions with voice commands (e.g., "Start new chat")
Voice Shortcuts Custom voice shortcuts for frequently used actions
Voice UI Accessibility Voice-first interface for users who rely on spoken interaction

Voice Training#

Feature Description
Custom Voice Training Train Iris to respond with a voice that matches your preference
Voice Profile Selection Choose from multiple voice profiles and styles
Voice Speed Control Adjust the speaking rate of AI responses
Voice Tone Adaptation AI adjusts tone (formal, casual, empathetic) based on context

Vision and Multimodal#

Iris can see, hear, and interact with the physical world.

Vision Features#

Feature Description
Image Analysis Upload images for AI analysis, description, and Q&A
Screenshot Analysis Take screenshots for the AI to analyze and respond to
Document Scanning Scan paper documents with your camera for AI processing
Diagram Understanding AI understands charts, diagrams, and visual data
OCR Extract text from images of documents, signs, and handwriting

Spatial Interaction (XR)#

Feature Description
Spatial AI Assistant Interact with Iris in 3D space using VR/AR headsets
Gesture Control Control Iris with hand gestures in XR environments
Spatial Pinning Pin AI responses to physical locations in AR
3D Object Recognition AI identifies and discusses objects in your physical environment

IoT Integration#

Feature Description
Smart Device Control Control IoT devices through natural language with Iris
Sensor Data Analysis AI analyzes data from connected sensors
Ambient Intelligence Iris responds to environmental context from connected devices

Brain-Computer Interface (BCI)#

Feature Description
BCI Input Experimental support for brain-computer interface devices
Neural Signal Processing Process neural signals for intent recognition
Adaptive Responses AI adapts to cognitive state detected via BCI

Emotional Intelligence#

Iris understands and responds to emotional context.

Emotional Features#

Feature Description
Emotion Recognition Detect emotional cues from text, voice tone, and context
Empathetic Responses AI responds with appropriate empathy and sensitivity
Rapport Building Iris builds conversational rapport over time
Social Context Awareness Understands social dynamics and adjusts communication style
Wellbeing Monitoring Gently monitors user wellbeing and offers support if needed
Ethical Reasoning AI applies ethical reasoning in sensitive conversations
Mood Tracking Track emotional patterns over time with user consent

Style Adaptation#

Feature Description
Tone Adjustment Switch between formal, casual, professional, and friendly
Verbosity Control Request concise or detailed responses
Format Preferences Prefer bullet points, paragraphs, tables, or code blocks
Audience Awareness Adjust complexity for beginners, experts, or executives
Cultural Sensitivity Adapt communication style for different cultural contexts

Reasoning and Thinking#

Advanced reasoning capabilities for complex problems.

Reasoning Features#

Feature Description
Chain-of-Thought AI shows its reasoning step by step for transparency
Extended Thinking Deeper reasoning for complex problems that need more thought
Self-Consistency Checks AI verifies its own reasoning for internal consistency
Metacognitive Monitoring AI monitors its own confidence and flags areas of uncertainty
Uncertainty Disclosure Clear communication when the AI is uncertain about an answer
Confidence Scoring Each response includes a confidence indicator

Agent Marketplace#

Discover and install specialized AI agents and plugins.

Marketplace Features#

Feature Description
Agent Directory Browse a curated directory of specialized AI agents
Plugin Library Install plugins to extend Iris's capabilities
One-Click Install Install agents and plugins with a single click
Reviews and Ratings Community reviews to help choose the best tools
Developer Publishing Developers can publish their own agents and plugins
Version Management Automatic updates with rollback capability

Agent Archetypes and Personalities#

Feature Description
Agent Archetypes Pre-defined agent templates for common use cases
Custom Personalities Customize agent personality, tone, and behavior
Agent Specialization Fine-tune agents for domain-specific expertise
Personality Profiles Choose from different personality styles for your AI assistant

Privacy and Security#

Iris is built with privacy as a core principle.

Privacy Controls#

Feature Description
End-to-End Encryption Conversations encrypted so only you can read them
Local-Only Mode Run AI models entirely on your device with no cloud data
Data Minimization Collect only what is necessary to provide the service
Differential Privacy Analytics that mathematically protect individual privacy
Consent Management Granular control over all data processing activities
Data Residency Choose which geographic region stores your data
Data Portability Export all your data at any time in standard formats
Data Deletion Delete all personal data and conversations on request
Audit Logging Complete record of all data access and processing

Security Features#

Feature Description
API Key Management Generate, rotate, and revoke API keys from the dashboard
Rate Limiting Per-user and per-IP rate limiting to prevent abuse
Secret Detection Automatic detection and redaction of sensitive data in inputs
Sandboxed Execution All tool and code execution runs in isolated sandboxes
Permission System Fine-grained permissions for what agents can access
Access Control Role-based access control for enterprise deployments

Anonymization#

Feature Description
Data Anonymization Strip personally identifiable information from analytics
Anonymous Conversations Option for conversations that are not logged or stored
Privacy Dashboard Centralized view of all privacy settings and data usage

Compliance#

Standard Description
GDPR Full GDPR compliance with data access, deletion, and portability
CCPA California Consumer Privacy Act compliance
SOC 2 SOC 2 Type II security certification
ISO 27001 International information security management standard

Accessibility#

Iris is designed to be accessible to everyone.

Visual Accessibility#

Feature Description
Screen Reader Support Full compatibility with VoiceOver, NVDA, JAWS, and TalkBack
High Contrast Mode Enhanced contrast themes for low vision users
Large Text Support Scalable text size across the entire interface
Color Blind Modes Color themes for protanopia, deuteranopia, and tritanopia
Braille Display Support Integration with refreshable braille displays

Motor Accessibility#

Feature Description
Full Keyboard Navigation Complete keyboard-only navigation throughout the app
Switch Access Support for single-switch and scanning input methods
Voice Control Navigate and interact entirely through voice commands
Dwell Control Activate elements by hovering for a configurable duration
Custom Shortcuts Define custom keyboard shortcuts for frequent actions

Hearing Accessibility#

Feature Description
Closed Captions Real-time captioning of AI voice responses
Visual Notifications Visual alerts instead of audio notifications
Transcript Generation Full text transcripts of voice conversations

Cognitive Accessibility#

Feature Description
Simplified Interface Option for a simplified, less cluttered interface
Reading Level Adjustment Request AI responses at a specific reading level
Step-by-Step Mode Break complex responses into manageable steps
Summary Mode Get brief summaries instead of long-form responses
Reduced Motion Disable animations for motion-sensitive users

Internationalization#

Feature Description
100+ Languages AI conversation support in over 100 languages
RTL Support Full right-to-left language support
Localized Interface Interface translated into major world languages
Auto-Language Detection Automatic detection of the user's preferred language

Personalization#

Iris adapts to you over time.

Personalization Features#

Feature Description
Communication Preferences Iris learns your preferred response style, length, and format
Domain Expertise Modeling Iris adjusts complexity based on your expertise in each topic
Tool Preferences Remembers which tools and workflows you prefer
Schedule Awareness Considers time of day and your typical patterns
Context Carry-Over Important context carries across conversations automatically
Proactive Suggestions AI proactively offers relevant help based on your patterns
Memory-Based Personalization Responses improve over time as Iris learns about you

Analytics and Insights#

Understand how you use AI and track your productivity.

Usage Analytics#

Feature Description
Conversation Metrics Total conversations, messages, and time spent
Model Usage Breakdown of which AI models you use most
Token Consumption Track token usage across models and conversations
Response Quality Rate responses and track quality trends over time
Usage Patterns Insights into when and how you use Iris most

Productivity Insights#

Feature Description
Time Saved Estimates Estimated time saved by using AI assistance
Task Completion Rates Track how often AI-assisted tasks are completed
Knowledge Utilization How effectively your knowledge base is being used
Agent Performance Which agents are most effective for your use cases

Notifications and Presence#

Notifications#

Feature Description
Task Completion Alerts Notified when background agent tasks finish
Knowledge Updates Alerts when documents in your knowledge base are updated
System Announcements Platform updates, new features, and maintenance notices
Configurable Channels Choose notification delivery via push, email, or in-app

Presence#

Feature Description
Cross-Device Sync Conversation state syncs across all your devices in real time
Presence Indicators See which devices have Iris active
Handoff Between Devices Seamlessly continue a conversation on a different device

Platform Administration#

For enterprise and team deployments.

Administration Features#

Feature Description
User Management Create, manage, and deactivate user accounts
Role-Based Access Control Define roles with specific permissions and capabilities
Usage Quotas Set token and request limits per user, team, or organization
Billing Management View and manage subscription plans and billing
Model Configuration Configure which AI models are available to users
Compliance Settings Configure data retention, privacy, and compliance policies
Custom Branding White-label Iris with your organization's branding

Platform Streaming#

Feature Description
SSE Streaming Server-Sent Events for real-time response delivery
WebSocket Streaming WebSocket connections for bidirectional real-time communication
gRPC Streaming High-performance streaming for internal service communication

Rate Limiting#

Feature Description
Per-User Limits Configurable request limits per user
Per-IP Limits IP-based rate limiting to prevent abuse
Token Budgets Monthly token budgets with alerts and enforcement
Burst Allowances Allow temporary bursts above the normal rate limit

Webhooks#

Feature Description
Event Webhooks Receive HTTP callbacks for conversation events
Configurable Events Choose which events trigger webhook notifications
Webhook Authentication HMAC signature verification for webhook security
Retry Logic Automatic retry with exponential backoff for failed deliveries

Integrations#

Iris integrates with other Oshun domains and external services.

Oshun Domain Integrations#

Domain Integration
Nyx Ask Iris about astronomical events, star positions, and more
Sophia Connect Iris to research knowledge bases and academic sources
Psyche Integration with the AI virtual assistant platform
Maya AI assistance for game development and world building
Yemaya Creative AI tools for art, music, and storytelling
Hathor World-building AI assistance with lore and narrative generation

External Integrations#

Feature Description
Plugin Architecture Extensible plugin system for adding new capabilities
Webhook Support Connect Iris events to external services and workflows
API Gateway Unified API gateway for all Iris services
Third-Party Tools Integration with popular tools (Slack, Notion, Jira, etc.)

Developer Platform#

Build on top of Iris with comprehensive developer tools.

SDK#

SDK Language Description
@iris/sdk TypeScript Full-featured TypeScript SDK for all Iris APIs
iris-sdk Python Python SDK with async support

API Access#

Method Use Case Latency
REST CRUD operations, batch processing < 50ms
WebSocket Real-time streaming, events < 20ms
gRPC High-performance internal calls < 10ms

API Endpoints#

Endpoint Description
GET /api/v1/conversations List all conversations
POST /api/v1/conversations Create a new conversation
GET /api/v1/conversations/:id Get conversation details
DELETE /api/v1/conversations/:id Delete a conversation
POST /api/v1/conversations/:id/messages Send a message and receive AI response
GET /api/v1/conversations/:id/messages Get message history
POST /api/v1/conversations/:id/stream Stream a response via SSE

Developer Tools#

Feature Description
Developer Portal Web-based documentation, API explorer, and key management
API Playground Interactive API testing environment
SDK Documentation Comprehensive SDK documentation with examples
SDK Testing Utilities Test utilities for mocking and testing Iris integrations
SDK Versioning Semantic versioning with backward compatibility guarantees
Code Generation Generate typed API clients from OpenAPI specifications
Custom Agent Creation Build and deploy custom AI agents
Plugin Development Develop plugins that extend Iris capabilities
Webhook Testing Tools for testing webhook integrations locally

Wearable and XR#

Wearable Features#

Feature Description
Watch App Quick AI interactions from your smartwatch
Voice-First Interface Speak to Iris from your wrist
Quick Replies Pre-set quick replies for common interactions
Haptic Notifications Feel notifications through haptic feedback
Glanceable Answers Brief, at-a-glance AI responses optimized for small screens

XR Features#

Feature Description
VR Assistant Full AI assistant experience in virtual reality
AR Overlay AI information overlaid on the real world
Spatial Conversations Conversations pinned to specific locations in 3D space
Object Recognition Point at objects for AI identification and information
Holographic Display AI responses rendered as holographic elements in AR
WebXR Compatibility Works in any WebXR-compatible browser and headset

Feature Summary#

Category Feature Count Highlights
Conversations 17 Streaming, editing, attachments, branching, export
AI Models 12 6+ providers, auto-routing, ensemble, fallover
Memory System 22 4-tier hierarchy, consolidation, visualization, export
Knowledge Base 15 Semantic search, citations, graph, fact checking
AI Agents 18 9+ built-in agents, tools, A2A, MCP, multi-agent
Code Intelligence 12 Generation, review, debugging, codebase analysis
Voice 12 Input, output, wake word, commands, custom training
Vision & Multimodal 12 Image analysis, XR, IoT, BCI
Emotional Intelligence 12 Recognition, empathy, rapport, wellbeing monitoring
Privacy & Security 18 E2E encryption, local mode, GDPR, SOC 2, secret detection
Accessibility 18 Visual, motor, hearing, cognitive, 100+ languages
Developer Platform 15 SDKs, APIs, marketplace, plugins, webhooks
Wearable & XR 11 Watch app, VR/AR assistant, spatial conversations

Iris: Intelligence, everywhere.