MiAiUi
Run — and watch — your Claude agents

Everything you build with Claude, in one place.

MiAiUi turns the skills, agents, and solutions you've built with Claude into a visible, searchable control panel — run them with one click and watch them work live on a map of what they're thinking, touching, and spawning. No terminal required.

A companion to Claude Code — the claude CLI stays the engine; MiAiUi is the cockpit.

Stop hunting for what you've built

Skills and automations pile up fast, and most of them live as files you have to remember and commands you have to retype. MiAiUi indexes all of it and gives you a real interface — so the work you've already done is actually usable, every day.

search skills… 282 skills All ★ Favorites writing ops
Every skill and plugin, indexed and searchable — one click to run, no names to remember.
The part everyone remembers

Watch your agents actually work

Most tools treat an AI run as a black box — a prompt goes in, an answer comes out. MiAiUi opens it up. Every run becomes a live Activity Map you can watch, and replay.

6Files 3Search 4Commands 2Web 1Knowledge ⬢Sub-agent Agent working $0.1306 · 48k tok
The live Activity Map — the agent at the center, everything it touches lighting up around it.
🧠

Its mind

Watch the agent's plan check itself off and its thinking stream in real time — intent, not just output.

🛰️

Its reach

A living graph with the agent at the center and a node for everything it touches — files, commands, web — lighting up as it works.

🤝

Its team

When an agent spawns sub-agents, watch the whole team collaborate — each with its own tool calls nested underneath.

⏪

Replay it

Scrub back through any finished run and watch it rebuild moment by moment, with live cost and tokens.

What MiAiUi does

One cockpit for discovery, running, monitoring, and scheduling.

🗂️

Unified catalog

Every skill and plugin, indexed and searchable with rich cards, categories, and favorites — no more remembering names.

🤖

Agents

Browse every agent available to Claude as rich cards, and create your own — name, role, tools, and system prompt — right in the app.

▶️

One-click run

Launch any skill with a click. Skills with inputs get a generated form; the rest take a simple prompt.

🗺️

Live agent map

Every run as a three-pane map — mind, reach, and cost — with a multi-agent mesh, flowing activity, and full replay.

🧭

Multi-run dashboard

Every active run as a live tile: status, current step, elapsed, cost. Click any tile for the full monitor.

🧩

Solutions

Chain skills into an ordered bundle and run them in sequence, with per-step prompts and live progress.

⏰

Scheduling

Run a skill or solution on an interval or daily at a set time, with weekday selection and run-now.

📊

Usage analytics

Totals, cost-per-day trends, and breakdowns by skill and model — know exactly what you're spending.

🩺

Doctor

Flags the problems that stop a skill from being discovered — missing names, bad frontmatter, collisions.

✦

Simple mode

A no-jargon view — just a library and an activity feed — so anyone on your team can run your tools.

How it works

From scattered files to a working cockpit in three steps.

Scheduled · every 1h ⚡ Triggerwebhook / cron Skill 1research Skill 2draft + send ✓ on successnotify + log ✕ on errorrun handler
Chain skills into a solution, branch on success or error, and put it on a schedule — set once, runs itself.

Point it at your work

MiAiUi scans your Claude skills, plugins, and scheduled tasks and builds a unified, searchable catalog.

Run & watch

Click to run, and watch it come alive on the Activity Map — its plan, its thinking, every file and command it touches, and any sub-agents it spawns.

Automate it

Bundle skills into solutions and put them on a schedule, so the work you do once keeps happening.

Why teams use it

Built to make the AI work you've already done pay off.

Cost / day last 7 days MTWTFSS This week $12.40 runs34 tokens1.2M avg / run$0.36 ● every run tracked
Every run records cost and tokens — totals, daily trends, and breakdowns by skill and model.
✓
See inside every runNo black-box AI — watch what the agent thinks, touches, and spawns, and replay it later.
✓
Your agents, organizedBrowse all your agents and create new ones without hand-editing a single file.
✓
No terminal requiredSimple mode makes your skills runnable by people who never touch a command line.
✓
Self-hosted & privateRuns on your own machine against your own Claude account — your data stays local.
✓
Cost in the openEvery run records cost and tokens, so spend is never a mystery.
✓
A companion, not a replacementClaude Code stays the engine; MiAiUi is the interface on top of it.
✓
Reachable anywhereOpen it from your phone or any browser, secured behind access control.
✓
Nothing left behindFull run history with replayable transcripts for everything you've run.

Ready to see your work in one place?

Open the app and your catalog is waiting.