v1.0.52 — May 20, 2026
Usage Analytics Dashboard
- Richer token insights — the Tokens chart now breaks down usage by input, output, and routing in a stacked view with distinct colors, making it easy to see where your AI spending goes
- Custom tool activity — a dedicated series tracks custom tool runs separately from built-in tools, so you can see how often your integrations are actually used
- More accurate KPIs — dashboard metrics now correctly attribute input vs. output tokens, filter to the right scopes, and align totals across all views
Conversation Experience
- Message-level metrics — model, provider, and token counts are tracked per message and visible in the dashboard
- Metrics in the message footer — conversation statistics moved to the input footer for a cleaner, always-visible layout
Reliability and Performance
- System tray initialization is now idempotent — no duplicate tray icons on restart
- Soft-delete archiving with a configurable TTL keeps your conversation history tidy and recoverable
- The app now waits for services to be fully ready before acting on scheduled prompts and health checks
Account and Subscriptions
- Sign-out flow now shows a clear warning if your session was revoked from another device
- Subscription state refreshes immediately after billing events — no manual reload needed
- Continued security and reliability enhancements to the authentication and commerce subsystems
v1.0.51 — May 13, 2026
Account and Subscriptions
- Free and paid tiers — sign in to unlock higher usage limits, or continue anonymously with the starter tier
- Usage at a glance — your current usage, plan tier, and weekly reset countdown appear in the account flyout
- Seamless checkout — subscribing and managing your subscription open directly in your browser; the app refreshes your entitlement immediately after checkout
- Promo codes — promo codes are now accepted at checkout
- Automatic usage sync — usage counts sync when you reconnect to the internet, resume from sleep, or bring the app to the foreground
Terms of Service and Privacy Policy
- Prosponsive now includes a Terms of Service, EULA, and Privacy Policy, presented on first launch
- Anonymous users can accept before signing in, and their acceptance carries over when they create an account
Multi-Agent Orchestration
- Coordinated agent chains — AI chains can now coordinate multiple specialized agents in a single response; maximum chain depth is configurable in Settings
- Configurable limits — the maximum number of tool steps per response and the AI inference timeout are now configurable in Settings
Tool Approval and Auto-Approval Rules
- Confidence badges — tool approval cards show the AI model's self-rated confidence as a color-coded badge
- Confidence thresholds — auto-approval rules support an optional confidence threshold: auto-approve only when confidence meets your threshold, or auto-deny when it falls below
- Approvals survive restarts — pending tool approvals are preserved across app restarts
- Tool categories are now labeled "Core tools" and "Custom tools"
Thread Summaries
- Automatic summaries — collapsed threads show a conversation summary; if the model skips generating one, the app retries in the background
- Persistent collapse state — thread collapse/expand state is saved across sessions
- The thread summary prompt is customizable in Settings
Windows
- Code-signed builds — Windows builds are now code-signed via Azure Trusted Signing, eliminating the SmartScreen warning on launch
Reliability
- Conversation messages that fail to save due to a transient database error are now retried automatically
- Custom tool discovery retries automatically when the initial cache is empty after a failed startup
- Fixed several edge cases in session restore, token refresh, and sign-in state after restart
- Fixed completed tool calls occasionally re-executing on re-entry into a thread
- Default Ollama model is now llama3.1:8b
- The local API server now requires authentication, so only Prosponsive's own services can reach it
Security
- Electron renderer runs in a sandboxed process with a Content Security Policy
- All TLS connections enforce TLS 1.2 minimum with FIPS-approved ciphers
- Container bind mounts are read-only where possible; log files are owner-readable only
- Auto-update downloads are verified with SHA-512 checksums and publisher signature before installation
v1.0.50 — April 17, 2026
Security
- Continued security enhancements across the platform
More Reliable Installation
- Smoother setup experience — Prosponsive guides you through each step more reliably, confirming services are ready before moving forward
- Automatic recovery — if something gets out of sync during setup, Prosponsive detects and repairs it automatically
Smarter Daily Use
- Cleaner tool names — tools show simple names like "Archive Channel" instead of cluttered internal prefixes
- Dates in your timezone and language — timestamps are formatted for your locale, not displayed as raw UTC
- Automatic provider failover — if your AI provider goes down or you hit a rate limit, Prosponsive switches to your next provider and keeps working
Better Visibility
- Improved install analytics — helps us identify and fix setup issues faster for new users
Platform
- Windows code signing — Windows builds now provide a verified installation experience
v1.0.49 — April 15, 2026
- Completely redesigned install and onboarding experience — a cleaner, faster path to getting started with Prosponsive
- SSL certificate trust now works seamlessly — no more browser security warnings or password prompts
- Inline replies in conversations are now perfectly positioned — no overlapping, shifting, or clipping
- Fixed a macOS Dock bug that could show duplicate Prosponsive icons
- Guest agent delegation is now more reliable — thread replies work correctly in shared channels
- Switched to PostHog analytics for better insight into how we can improve the app
- Chat state is more robust — eliminated optimistic update glitches and cache drift issues
- AI provider connections now require your explicit approval before activating, giving you full control
v1.0.47 — April 9, 2026
- AI Provider configuration fixes
- Your custom n8n workflows are now preserved when Prosponsive updates — no more losing your changes on restart
- Performance dashboard now correctly attributes tool usage between your custom tools and built-in Prosponsive tools
- Fixed first-run provider setup so you can proceed after connecting your AI provider
v1.0.41 — April 7, 2026
- Now available in 12 languages — Spanish, Portuguese, French, German, Japanese, Chinese, Korean, Italian, Indonesian, Thai, and Vietnamese join English
- Dates and times adapt to your language — month names, day labels, and clock format all follow your locale
- Automatically matches your OS language — or pick any supported language manually
- More reliable startup — Prosponsive now waits for your database to be fully ready before connecting
- Smoother experience with Podman 1.5+ — network setup happens automatically so containers start on the first try
- Agents can always hand off to the right specialist — delegation back to the default agent works reliably
- Clearer @mention dropdown — selected items stand out with better contrast
- New Performance page — see how your tools and agents are being used at a glance
- Stay informed with "What's new" — the update banner now links directly to these release notes
- Introducing the Notify tool — agents can send you native notifications with one-click navigation back into the conversation
v1.0.38 — April 6, 2026
- Schedule your AI to work while you're away — set up one-time or recurring prompts that run automatically
- Quickly tag agents and channels with @-mentions that autocomplete as you type
- Get more out of every conversation — Prosponsive now automatically adapts to each model's full context window
- New AI providers are ready to use the moment they're discovered — no manual setup needed
- Smoother experience switching between conversations — agents no longer lose their place
- More reliable first-launch setup on Windows with Docker Desktop
v1.0.37 — April 3, 2026
- Introducing Prosponsive — your proactive, responsive AI desktop app that works alongside you with local and cloud AI models
- Connect to the AI that works best for you — Ollama, OpenAI, Anthropic, Google, AWS Bedrock, Groq, and Mistral all supported out of the box
- Never lose a conversation — threaded channels keep your work organized with drag-to-reorder, archiving, and cross-channel thread moves
- Put your AI on autopilot — auto-approve trusted tools or review each one before it runs, with full visibility into what your agents are doing
- Stay resilient when things go wrong — automatic failover switches to your next-best AI provider with zero disruption
- Build your own agent team — create custom agents with specialized tools, share them to specific channels, and let them delegate to each other
- Works on Mac and Windows — with automatic updates so you always have the latest features
- Runs Docker or Podman — Prosponsive detects and configures your container runtime automatically, including SSL certificates you can trust with one click
