The Roadmap
What's coming next
Phase 1 is complete.
Recall, Reflex, Pulse, Signal—the core observability stack is live. Vault for secrets. Flux for custom metrics. Dendrite for documentation. Vision for visual testing. Synapse for AI command. Platform tying it all together.
Now for Phase 2.
Beacon: Uptime Monitoring
Your app is running. But is it up?
Beacon monitors endpoints from multiple global locations:
- NYC, London, Singapore — Check from where your users are
- HTTP, SSL, DNS, TCP — Different check types for different needs
- SSL expiry tracking — Know before certificates expire
- Public status pages — Show users you're transparent
If your API goes down at 3am, Beacon notices. Signal alerts you. You fix it before users complain.
ETA: February 2025
Cortex: Feature Flags
Ship features safely:
- Boolean flags — On/off for everyone
- Percentage rollouts — 10% of users, then 50%, then 100%
- Segment targeting — Pro users only, beta testers only
- A/B tests — Which version performs better?
if Cortex.enabled?(:new_checkout, user: current_user)
render_new_checkout
else
render_old_checkout
end
Redis-backed for speed. ActionCable for real-time updates. Kill switches when things go wrong.
ETA: February 2025
Nerve: Job Monitoring
Background jobs are invisible until they fail.
Nerve watches them:
- Sidekiq, Solid Queue, Resque, DelayedJob, GoodJob — Whatever you use
- Queue health — Depth, latency, throughput
- Job metrics — Duration, success rate, retry rate
- Dead job morgue — Investigate and retry failed jobs
"Why is the queue backed up?" becomes answerable.
ETA: March 2025
Sentinel: Host Monitoring
Applications run on hosts. Hosts have problems too.
Sentinel monitors infrastructure:
- CPU, memory, disk, network — The basics
- Process monitoring — Is nginx running?
- Container stats — Docker metrics
- Lightweight agent — ~10MB Go binary, minimal overhead
A Go agent that reports to Signal. Full infrastructure visibility.
ETA: March 2025
The Phase 2 vision
After Phase 2, you'll have:
| Layer | Product | Status |
|-------|---------|--------|
| Infrastructure | Sentinel | Coming |
| Uptime | Beacon | Coming |
| Feature Flags | Cortex | Coming |
| Jobs | Nerve | Coming |
| Logs | Recall | Live |
| Errors | Reflex | Live |
| Traces | Pulse | Live |
| Metrics | Flux | Live |
| Alerts | Signal | Live |
| Secrets | Vault | Live |
| Docs | Dendrite | Live |
| Testing | Vision | Live |
| AI Command | Synapse | Live |
| Platform | Platform | Live |
Complete observability. Complete automation. One stack.
Phase 3: Intelligence
Phase 1 was collection. Phase 2 is coverage.
Phase 3 is intelligence:
- Automated incident response — AI that fixes problems, not just reports them
- Predictive alerts — Know about issues before they happen
- Cost optimization — Suggestions for reducing infrastructure spend
- Security analysis — Proactive vulnerability detection
The stack becomes proactive. Not just watching—acting.
ETA: Q3 2025
Phase 4: The AI Company
The end goal hasn't changed.
An AI-run company. Humans set direction. AI handles execution.
Synapse coordinates. Products provide data. Hive Core executes.
The stack isn't just for observability. It's the nervous system of an autonomous organization.
We're building the infrastructure for what comes next.
Follow along
Every step is documented here. Every decision explained.
Building in public means you see the roadmap evolve. Features ship. Plans change.
This is the journey.
— Andres