Styrby Now Supports 11 CLI Coding Agents with Session Checkpoints, Voice Commands, and Enterprise OTEL Export
The CLI coding agent space has expanded faster than most developers can keep up with. In the last year, the number of serious contenders went from a handful to more than a dozen, each with a different backend, pricing model, and workflow. Developers running multiple projects now routinely juggle three, four, or five agents at the same time. The overhead of monitoring all of them from separate terminals, tracking costs across separate billing dashboards, and approving permissions mid-session has become its own kind of work.
Styrby was built to remove that overhead. This post covers everything shipping in the current sprint: four new agent integrations bringing the total to 11, plus session checkpoints, session sharing, per-message cost tracking, voice commands, cloud task monitoring, mobile code review, activity graphs, OTEL metrics export, LiteLLM dynamic pricing, and a Rust-powered JSONL parser for large session files.
What Is New in This Sprint
Four New Agents (11 Total)
Styrby now connects to Crush, Kilo, Kiro, and Droid in addition to the seven agents already supported: Claude Code, Codex, Gemini CLI, Aider, OpenCode, Goose, and Amp. Every agent appears in the same unified dashboard with real-time status, session cost, and permission requests routed to your phone.
Session Checkpoints
Long-running agent sessions can now be checkpointed at any point. A checkpoint captures the full session state so you can restore it later, branch from it to try a different approach, or hand it off to a colleague. Checkpoints are stored encrypted in the cloud and appear in your session history alongside normal session entries.
Session Sharing with E2E Encryption
Any session can now be shared via a link. The link includes a separate encryption key that is never stored on Styrby servers. The recipient decrypts the session locally in their browser. Revoke the link at any time from your dashboard and the session becomes unreadable to anyone who has not already loaded it.
Per-Message Cost Tracking
Every message in a session now shows the exact cost of that exchange: input tokens, output tokens, cache hits, and the dollar amount at the model's current rate. You can see at a glance which prompts are expensive and where context window bloat is driving up costs, without waiting for the end-of-session summary.
Voice-to-Agent Commands
You can now speak commands to any connected agent from the Styrby mobile app. The app transcribes your voice and sends the command to the agent via the CLI bridge. Useful for quick instructions when your hands are occupied or when you are away from your desk and want to redirect a running session without typing on a small screen.
Cloud Task Monitoring
Agents that support async task execution (long-running background jobs) now surface task status in Styrby. You receive a push notification when the task completes, when it needs input, or when it encounters an error. The task log is available in the mobile app without opening a terminal.
Code Review from Mobile
Agents that produce diffs can now submit them to the Styrby review queue. You see the changed files, the diff, and the agent's explanation on your phone. Approve or reject from the notification or from the review screen in the app. The decision is relayed back to the agent immediately.
Activity Graph
The dashboard now includes a GitHub-style contribution heatmap showing session activity by day over the past year. Each cell represents one day; intensity reflects session count and total cost. It is a fast way to spot usage patterns, identify periods of high spend, and see how your agent usage correlates with project milestones.
OTEL Metrics Export
Styrby can now push session metrics to any OpenTelemetry-compatible backend: Grafana, Datadog, Honeycomb, New Relic, and anything else that accepts OTLP. Export includes session duration, token counts, cost, agent type, model, and custom tags. Configure the endpoint and headers once in settings and every session ships to your observability stack automatically.
LiteLLM Dynamic Pricing
Cost calculations now pull from the LiteLLM pricing database, which covers more than 300 models and updates when providers change their rates. You no longer need to update Styrby or wait for a release to get accurate cost estimates for a new model. If your agent switches models mid-session, each message is costed at the correct rate for the model that generated it.
Rust-Powered JSONL Parser
Session files from agents like Aider and OpenCode can grow large over long sessions. Parsing them in JavaScript became a bottleneck for sessions with tens of thousands of messages. The parser is now a native Rust module compiled to WebAssembly. Large session files that previously took several seconds to load now parse in under 200 milliseconds.
Agent Spotlight
Goose (Block / Square)
Goose is Block's open-source agent with enterprise Model Context Protocol support. MCP lets Goose connect to external data sources, internal APIs, and company-specific tools via a standardized protocol. Styrby monitors Goose sessions and surfaces MCP tool calls in the permission approval flow, so sensitive MCP actions still require explicit sign-off from your phone.
Amp (Sourcegraph)
Amp is Sourcegraph's agent built on their code intelligence platform. Its deep mode can spawn sub-agents to work on different parts of a codebase in parallel. Styrby tracks each sub-agent as a child session under the parent Amp session, so you see the full cost tree and can approve permissions for any branch of the parallel work without switching to the terminal.
Crush (Charmbracelet)
Crush is Charmbracelet's agent with their signature terminal UI polish. It runs in a Bubble Tea TUI that is genuinely pleasant to use. Styrby connects to Crush through the standard CLI bridge and mirrors session output to the mobile app, preserving the color-coded TUI output so the session is readable on a phone screen.
Kilo
Kilo routes to more than 500 models through a unified API layer and adds a Memory Bank feature that persists context across sessions. Styrby tracks Kilo's multi-model routing and correctly attributes each message to the model that handled it, so your cost breakdown reflects the actual model mix rather than a single blended rate.
Kiro (AWS)
Kiro is Amazon's IDE-integrated agent with a per-prompt credit system instead of token billing. Styrby maps Kiro's credit consumption to a dollar-equivalent estimate using the published credit rate, keeping the cost dashboard consistent across agents that use different billing models.
Droid
Droid is a bring-your-own-key agent that routes to multiple backend providers based on availability and cost. Styrby tracks which backend Droid selects for each request and breaks down costs by provider, giving you visibility into how the routing decisions affect your bill across different API keys.
Security
Session content is encrypted end-to-end using TweetNaCl box encryption. Keys are generated on your device and never sent to Styrby servers in plaintext. When you share a session, the share link includes a derived key that is separate from your primary session key. Revoking the share link does not compromise other sessions or your master key.
The full codebase went through an OWASP-guided security audit in Q1 2026. The audit covered authentication flows, RLS policy coverage on all 16 Supabase tables, rate limiting, webhook signature verification, and the CLI bridge protocol. Findings were remediated before this sprint shipped.
OTEL export credentials (endpoint URL and headers) are stored in Vercel environment variables and never appear in session data or logs. The export pipeline runs server-side; your observability credentials are not exposed to mobile clients.
Pricing
| Plan | Agents | Price | Includes |
|---|---|---|---|
| Free | 3 agents | $0 / month | Session monitoring, push notifications, cost tracking, mobile app |
| Pro | 8 agents | $24 / month | Everything in Free plus session checkpoints, session sharing, voice commands, activity graph |
| Power | All 11 agents | $49 / month | Everything in Pro plus OTEL export, cloud task monitoring, mobile code review, team dashboard, per-developer cost attribution |
The Free plan lets you connect Claude Code, Codex, and one other agent. It covers the most common setup for solo developers and has no time limit. Upgrade to Pro when you need more agent coverage or the productivity features. Upgrade to Power when you need OTEL export or are managing AI costs across a team.
Getting Started
Create a free account at styrbyapp.com/signup. Install the CLI bridge on any machine where your agents run, connect your first agent in under five minutes, and install the mobile app to start receiving push notifications. No credit card required for the Free plan.
If you are already on the Free plan and want to unlock checkpoints, session sharing, or voice commands, upgrade to Pro from the billing settings in your dashboard. Power tier OTEL export requires a one-time endpoint configuration in settings; documentation is at styrbyapp.com/docs/dashboard.
Ready to manage your AI agents from one place?
Styrby gives you cost tracking, remote permissions, and session replay across 11 CLI coding agents.