Changelog
Launch Day
Stomme AI is live. The managed AI agent service is now available. Sign up, get onboarded, and your agent team starts running on your hardware.
✨ What's included
- 31 production-tested skills — Email triage, calendar management, research, content drafting, code review, file organisation, workflow automation, and more
- Three subscription tiers — Personal (€45.99/mo), Professional (€149.99/mo), Business (€599/mo)
- Founding offer — 50% off your first month (valid until 31 July 2026)
- Portal — Customer account management, usage dashboard, subscription controls
- Telegram integration — Talk to your agent from any device
- Three languages — Full website and portal in English, Swedish, and German
Portal & Onboarding
Customer portal launched.
- Account creation with magic link authentication
- Stripe-powered subscription management
- Self-service Telegram bot setup wizard
- Usage dashboard showing token consumption and agent activity
- Help centre with 15+ articles covering setup, billing, privacy, and troubleshooting
- Cancellation flow with downgrade and pause options
Pricing Confirmed
Three-tier pricing structure finalised.
- Usage caps set per tier (Claude tokens, ChatGPT tokens, web searches)
- Founding discounts confirmed: Personal 50% off, Professional 33% off, Business 34% off
- Currency support: EUR, SEK, USD, GBP — prices shown in visitor's local currency
- Cached tokens excluded from monthly usage caps
Skill Library Complete
31 skills passed adversarial security review.
Every skill in the library was generated, then attacked by an adversarial reviewer (the forge), rebuilt, and tested again. 3,500+ tests across the full library. Zero known vulnerabilities at launch.
Skills include: email infrastructure, calendar management, web research, content drafting, code review, file management, agent coordination, customer health monitoring, usage enforcement, and more.
v1.0.0 — Technical Details
✨ Core Platform
- Agent provisioning on macOS 13+ (Apple Silicon and Intel)
- Skill library with 31 production-grade skills, 3,500+ tests
- Adversarial forge pipeline — every skill generated, attacked, rebuilt, and hardened
- Mission Control dashboard for agent monitoring and configuration
- Full audit log — every agent action is recorded and reviewable
🔧 Skills (Selection)
- saas-scaffold — Core config, rate limiting, validation, audit logging
- saas-auth — HMAC sessions, magic links, GDPR compliance gates
- saas-email — SMTP client, templates, lifecycle emails
- saas-users — Account management, GDPR data export and deletion
- saas-payments — Stripe integration, subscription state machine, webhooks
- agent-architect — Agent persona, memory, approval gates, capability routing
✨ Mission Control
- Agent status dashboard with live heartbeat monitoring
- Skill and permission management UI
- Conversation history viewer with full audit trail
🔒 Privacy & Security
- All agent memory, files, and conversations stored locally on your computer
- Anthropic Claude API used under strict no-training data terms
- Path traversal protection on all file I/O operations
- High-impact actions require explicit customer approval before executing
✨ Communication Channels
- Telegram integration — send and receive from your agent via Telegram
- Email — agent reads and responds to your email via SMTP
- Webhooks — connect your agent to external services
✨ Tiers
- Personal — €45.99/month (€22.49 founding offer). 1 agent, full skill library, email support.
- Professional — €149.99/month. Up to 4 agents, code review, priority support.
- Business — €599/month. Up to 8 agents, team coordination, dedicated support.
🚀 Onboarding
- 45-minute personal onboarding call — we configure your agent for your workflow
- Automated installer: macOS provisioning script, agent configuration, and first run in under an hour
How to Read This Changelog
Each entry is tagged with the type of change:
- 🚀 Launch — Major release or product milestone
- ✨ New — New feature or capability
- 🔧 Improved — Enhancement to existing functionality
- 🐛 Fixed — Bug fix
- 🔒 Security — Security improvement or patch
We ship continuously. This log is updated with every meaningful release. Minor patches (dependency updates, typo fixes, internal refactors) are not listed.