Flagship cross-model interface

Valence Roadmap

Delivered foundation, free 1.x updates, longer-horizon vision.

From foundation and HIVE integration through kids mode, automations, imports, and premium polish, Valence already covers a broad surface area. The roadmap keeps shipped work, planned free updates, and future platform bets easy to scan.

65+
Features delivered
7
LLM providers
10+
Themes
30+
Keyboard shortcuts

Shipped & live

Everything shipped in Valence so far, from the Q1 foundation through post-launch delivery, local memory, automations, and kids mode.

Complete

Q1 2026

Foundation & Polish

Core Architecture

  • SRP Architecture: LEGO pattern with single-responsibility bricks
  • WebView2 UI: Modern HTML/CSS/JS interface
  • Multi-Provider: OpenAI, Anthropic, Google, Azure, xAI, Ollama
  • Ability System: #time, #calculate, #recall with progressive disclosure
  • Voice Triggers: “Hey Grok” and “Hey Claude” with fuzzy matching
6
Providers
5
Stability rounds

Complete

Q2 2026

HIVE Integration

Intelligent Memory

  • Memory Abilities: #recall and #remember for persistent AI memory
  • Auto-Recall: Natural language trigger detection with relevance scoring
  • Observable UX: Wave animations, relevance percentage, and source attribution
  • Cartridge System: Per-message and per-profile cartridge overrides
20+
HIVE features
100%
Configurable

Complete

Q3 2026

Progressive Disclosure

Context Intelligence

  • Tiered Prompts: Light initial prompts, full guides on first use
  • Auto-Summarization: Model-aware context window management
  • Token Tracking: Real-time usage visualization with before/after comparison
  • Intent Detection: Pre-emptive ability disclosure based on conversation
Context intelligence

The app now spends prompt budget more carefully by deciding when to surface full guidance, summaries, and ability hints.

Complete

Q4 2026

Premium Experiences

Polish & Power

  • Conversation Branching: Fork, edit, and regenerate from any point
  • Theme System: Multiple themes, font customization, compact mode
  • Voice Input: Push-to-talk, partial results, global hotkey
  • Multi-Modal: File attachments, image understanding, screenshot capture
  • Keyboard Shortcuts: Comprehensive system with categorized modal
25+
Features
ARIA
Accessible

Complete

Late 2026

Post-Launch Delivery

Shipped After v1.0

  • MCP Integration: Stdio transport, tool discovery, and auto-registered abilities
  • OpenRouter: 100+ models via a single API key
  • Multi-Tab: Create, close, and switch tabs with context sharing
  • Templates: Create, edit, delete, and insert reusable message templates
  • Licensing: 7-day trial, RSA key validation, and DPAPI encryption
  • First-Run Wizard: Guided setup, import/export, and auto-save recovery
11
Post-launch features
100+
OpenRouter models

Complete

v1.1 — March 2026

Local Memory & Import

Persistent Memory & Transfer

  • Local Memory: Persistent memory across conversations with LLM-powered summarization and tiered recall
  • ChatGPT Import: Import all conversations via browser login or ZIP export
  • Claude Import: Import conversations from Anthropic via browser login
  • Gemini Import: Import conversations from Google via browser login
  • Memory Browser: Search, browse, and manage saved memories
  • Bundled Web Search: Built-in MCP server for web search, auto-configured on first launch
3
Import providers
0
Setup required

Complete

v1.2–v1.3 — March 2026

Automations, Comparison & Polish

Automations, Comparison, and Refinement

  • Compare Models: Side-by-side comparison of two or three models in parallel
  • Scheduled Automations: Run prompts on a timer with file or command input sources
  • Image Generation: #image{prompt} with DALL-E, Imagen, Azure, and xAI support
  • API Usage Dashboard: Requests, tokens, and costs across providers
  • Window Context: Capture what’s on screen as AI context
  • Memory Facts: Structured fact extraction plus #forget for local memory
3
Models compared
4
Image providers

Complete

v1.4 — March 2026

Kids Mode & Parental Controls

Safety & Oversight

  • Kids Mode: Age-appropriate AI profiles for children ages 3–17 with five safety tiers and anti-jailbreak protections
  • Parent Dashboard: LLM-augmented conversation scanning with live progress, sentiment analysis, and crisis detection
  • Parental Controls: Password-protected lockdown, provider/model restriction, and daily session limits with extensions
  • Kids Wizard: Guided six-step setup with privacy consent, child profiles, safety previews, and icon picker
  • Session Timer: Per-profile daily tracking with countdowns, warnings, and parent-approved extensions
  • Crisis Detection: AI-powered analysis of distress signals with automatic crisis resource injection
5
Safety tiers
3–17
Age range

Free updates

Coming to all v1.x license holders at no additional cost. All 1.x bug fixes and minor updates stay free.

Free update

Auto-Save Controls

User-configurable auto-save frequency, save status in the UI, and multi-slot recovery so the last N drafts stay available instead of only the most recent.

Free update

Profile Quick-Switch

Switch profiles via #switch_profile, keyboard shortcuts like Ctrl+Alt+1–5, and fast profile duplication from the interface.

Free update

Conversation Tags & Search

Full-text search across all saved conversations plus lightweight tagging and filter-by-tag in the sidebar.

Free update

Template Variables

Support for placeholders like {DATE}, {TIME}, {PROFILE}, and {CLIPBOARD}, with starter templates for common workflows.

Free update

System Theme Detection

Auto-switch dark or light themes based on OS preference with a new System option alongside the existing theme picker.

Free update

Tab Pinning & Reorder

Drag-to-reorder tabs, pin important ones so they stay leftmost, and protect them from accidental closure.

Free update

Customizable Shortcuts

Rebind every keyboard shortcut from Settings, detect conflicts, and reset to defaults when needed.

Free update

Full Settings Backup

One-file export and import that includes profiles, templates, and all settings.

Free update

Provider Usage Stats

Track tokens per message, session cost estimates, and provider-level usage in a dedicated stats panel.

Free update

Notification History

Scrollable history of past notifications with per-type controls and do-not-disturb support.

The future

Stretch goals and larger version-shaping bets for agentic work, extensibility, collaboration, and governance.

Vision

AI that acts

Agentic Execution

AI That Acts

  • Task Execution: Multi-step workflows with tool orchestration
  • Tool Expansion: Beyond MCP into native filesystem, shell, and web tools
  • Plan & Execute: AI plans, user approves, Valence executes
Vision

Move from chat responses to practical, reviewable execution loops without hiding what the system is doing.

Vision

Community extensibility

Plugin Ecosystem

Community Extensibility

  • Ability SDK: Create and distribute custom abilities
  • Marketplace: Community ability sharing and discovery
  • Plugin System: Extend Valence with third-party integrations
Vision

Let the ecosystem expand the product surface while keeping the core interface calm and coherent.

Vision

Shared intelligence

Collaboration

Shared Intelligence

  • Multi-User HIVE: Team memory spaces with access control
  • Shared Conversations: Collaborative AI sessions
  • Platform Expansion: Mobile companion and browser extension
Vision

Bring shared memory and shared conversation into the product without sacrificing ownership or clarity.

Vision

Policy & compliance

Enterprise Governance

Policy & Compliance

  • Ability Safety Tiers: Safe, Network, Filesystem, Dangerous
  • Policy Engine: Machine-level policy files plus GPO and Registry support
  • Audit Logging: Who used what, when—full compliance trail
Vision

Make operator policy explicit and enforceable for environments that need safety tiers and traceable actions.