// request access

Join the Waitlist

Get console access the moment Opsdeck goes live.

No spam. Unsubscribe any time.

App 90 · 100 Apps 100 Hours

Mission control for your agent fleet.

Opsdeck is a command-center dashboard for monitoring AI agents. A live grid of agent tiles, throughput sparklines, action logs, and cost rings update in real time — so you catch a failed run the second it happens and pause or resume any agent from one console.

Book a Call
Watch the demo
Fleet Health
Live · 000

6

Active

1

Idle

1

Errored

throughput · req/s54.0
  • lead-scout
    active
  • enricher
    active
  • billing-recon
    error
  • deploy-sentinel
    idle

// the console

Five screens. One source of truth.

Fleet grid, agent console, activity stream, costs, and settings — every panel built for ops at a glance.

opsdeck01
Fleet Grid

/apps/90-agent-dashboard-ai/screens/01.png

01 / 05 · Fleet Grid

// live walkthrough

See the deck come alive

A 90-second tour from fleet overview to a paused agent and back.

opsdeck — demo.mp4

// capabilities

Run the fleet, not the fire drills

Everything updates live — no refresh, no guesswork.

home

Live Fleet Grid

Every agent as a live tile — status dot, current task, and a throughput sparkline that ticks in real time. Scan eight agents in one glance.

detail

Agent Console

Drill into any agent for a streaming action log, success-rate / latency / cost rings, and a pause-resume control that actually flips status and flattens throughput.

logs

Activity Stream

One filterable firehose of every run. Slice by level — ok, warn, error — or by agent, then deep-link any row straight to its console.

alerts

Alerts on Failure

The second a run fails, an alert card surfaces the agent and the error. Tap through to the console and act before it cascades.

costs

Cost & Usage

Today, 14-day, and average spend cards over a clean bar chart, plus per-agent cost bars and a model-routing tip that flags easy savings.

settings

Tuned to Your Ops

Set the refresh rate, panel density, and alert thresholds. The dial actually changes the tick speed — the whole deck speeds up or settles down.

Roadmap · Coming Soon

From dashboard to autopilot

Today Opsdeck shows you everything. Next, it acts — an AI ops co-pilot that triages alerts, predicts cost spikes, and re-routes work before a run ever fails.

Auto Alert Triage

AI reads the failed run, classifies the root cause, and drafts the fix — so an alert arrives with a diagnosis, not just a red dot.

Planned

Anomaly Detection

Models learn each agent's normal rhythm and flag drift in latency, cost, or throughput before it trips an alert.

Planned

Cost-Spike Forecasts

Project tomorrow's spend from today's run pattern and get a heads-up before a runaway agent burns the budget.

Planned

Smart Model Routing

An AI co-pilot suggests cheaper-model routes per task and can re-balance load across the fleet automatically.

Planned

// follow the build

Watch it ship in real time

Opsdeck is app #90 in the 100 Apps 100 Hours sprint. Every screen recorded, every build logged — follow along on socials for daily drops, ops notes, and early console access.

status: pre-launch

Running agents and flying blind?

Join the waitlist and be first on the deck when Opsdeck launches.

Bring your fleet online.

Join the waitlist for early console access, launch updates, and the first look at the AI ops co-pilot — straight to your inbox.

Book a Call

No spam. Unsubscribe any time. Prefer to talk? Grab a 30-minute slot.