Built for developers who ship with AI
End-to-end encrypted remote control for your AI coding agents. Cost tracking, permission approval, session history, and team collaboration.
Security and Encryption
Your code never touches our servers. Zero-knowledge by design.
End-to-End Encryption
Every session message is encrypted with TweetNaCl before it leaves your machine. Styrby relays ciphertext. We cannot read your prompts, responses, or code.
Public/private keypair generated locally. Keys never leave the device.
Zero-Knowledge Architecture
We process metadata only: timestamps, token counts, cost calculations, and connection status. Your intellectual property stays where it belongs.
Even if our database were breached, your code is unreadable ciphertext.
API Key Hashing
API keys are hashed with bcrypt before storage. We never store plaintext credentials. Rate limiting on all endpoints prevents brute-force attempts.
Per-endpoint rate limits with graduated backoff.
Audit Trail
Every permission approval, budget change, team action, and security event is logged with timestamp and actor. Exportable for compliance reviews.
Tamper-evident logging with row-level security.
Remote Agent Control
Approve, deny, and configure your AI agents from your phone or browser.
Permission Approval
Agents request permission before risky actions. You get a push notification with a risk badge (Low, Medium, High, Critical). Approve or deny in one tap.
Works from your phone, tablet, or any browser.
Control From Anywhere
Leave your desk. Go to lunch. Walk the dog. Your AI agents keep working, and you stay in control from your phone. No more babysitting a terminal.
Mobile app (iOS launching soon) plus responsive web dashboard.
Auto-Approve Rules
Configure rules for low-risk actions so agents proceed without waiting. Keep the approval gate on file deletions, shell commands, and network access.
Per-agent, per-tool granularity.
Blocked Tool Lists
Explicitly block dangerous tools per agent. Prevent rm -rf, database drops, or any command you define. The agent gets a clear denial instead of access.
Configurable per agent type.
Offline Command Queue
Lose connection? Commands queue locally and sync automatically when you reconnect. Your laptop can sleep. Styrby remembers.
Local queue with automatic sync on reconnection.
Cost Management
Know where every dollar goes. Set limits before you overspend.
Cost Tracking Across Every Agent
See spending per agent, per session, per model, and per tag. Daily trend charts, per-model breakdowns, and cost attribution across all five agents. Tag sessions by client or project to get a cost-by-tag breakdown for invoicing.
Materialized views for sub-second dashboard loads. Cost by Tag section shows total spend and session count per tag.
Model-Level Breakdown
Compare costs across Claude Opus, Sonnet, Haiku, GPT-4o, Gemini Pro, and every model your agents use. Input tokens, output tokens, cache hits.
Pricing reference table with last-verified dates.
Budget Alerts
Set daily, weekly, or monthly spend limits. Choose what happens when a threshold is hit: get notified, slow the agent down, or stop it entirely.
Per-agent or global. Pro gets 3 alerts, Power gets 5.
Multi-Agent Dashboard
Claude Code, Codex, Gemini CLI, OpenCode, and Aider in one view.
Unified Agent View
All five supported agents in a single dashboard. Color-coded status cards show active, idle, stuck, or failing states at a glance.
Real-time updates via Supabase Realtime subscriptions.
Error Attribution
When something breaks, know exactly what caused it. Color-coded sources: Orange (Styrby), Red (agent), Blue (build tools), Yellow (network).
Drill into error details from the dashboard.
Live Connection Status
Real-time heartbeat monitoring for every connected machine. Know instantly when a device goes offline or an agent stops responding.
WebSocket-based with automatic reconnection.
Session History
Full record of every agent conversation and action.
Searchable History
Filter sessions by agent type, tags, date range, or cost. Full-text search across session metadata. Bookmark important sessions for quick access.
Free: 7 days. Pro: 90 days. Power: 1 year.
Session Replay
Step through the full conversation between you and the agent. Every prompt, response, tool call, and permission request in chronological order.
End-to-end encrypted. Decrypted only on your device.
Per-Session Cost Breakdown
See exactly how much each session cost. Input tokens, output tokens, cache utilization, and total cost broken down by model used.
CSV export available on Power tier.
Notifications
Stay informed without being overwhelmed.
Push Notifications
Permission requests, budget alerts, error notifications, and session events delivered to your phone. React in seconds from anywhere.
APNs (iOS) and FCM (Android) supported.
Quiet Hours
Set time windows when notifications are silenced. Critical alerts like budget exceeded can optionally bypass quiet hours.
Per-notification-type granularity.
Weekly Summary Emails
Digest of your agent usage: total spend, active sessions, most-used agents, and budget status. Delivered every Monday.
Powered by Resend. Unsubscribe anytime.
Team Collaboration
PowerShare visibility across your engineering team.
Team Members
Invite up to 3 team members via email. Each gets their own login with role-based access: Owner, Admin, or Member. No shared credentials.
Invite flow with email verification.
Shared Dashboards
Team members see the same cost analytics, session history, and agent status. Engineering managers get visibility. Developers keep autonomy.
Row-level security ensures data isolation between teams.
REST API
Programmatic access to your session and cost data. Build custom integrations, internal dashboards, or CI/CD automations.
Authenticated with hashed API keys. Rate-limited.
Supported Agents
Ready to take control?
Free tier available. No credit card required. Connect your first agent in under two minutes.