Skip to content
Agent Selector
AgentSelector ME

Your life admin, handled — you stay the boss.

Build a personal agent that briefs you every morning, triages your inbox, and tracks what you care about — with approval gates, spend caps, and a kill switch, so nothing happens without your yes.

Build my personal agentFind my agent — 2-min quiz
Approval gates on sensitive actions Exact cost shown before every run Bring your own key — always free Kill switch on every agent

An agent that works for you — not on you

Autonomy with a leash. These four controls ship with every hosted agent on the platform.

Approval gates

Your agent drafts, sorts, and prepares — but a sensitive action (send, book, buy) waits for your explicit yes. Every time.

Hard spend caps

Set a budget and the agent stops at it. The exact cost is shown before every run — no surprise bills, ever.

Kill switch

One click halts any agent instantly. You’re never negotiating with your own automation.

Full audit trail

Every action your agent takes is logged and reviewable. If it did it, you can see it.

Four personal agents, ready to build

One click opens the Builder with the full brief — approval rules included — already written for you.

Morning Brief

Wake up to your day, already organized.

  • Delivers one morning digest: today's calendar, weather, and your priority emails
  • Surfaces news on only the topics you choose — no doomscroll feed
  • Flags schedule conflicts and suggests fixes before you're in them
  • Runs on a schedule, so it's waiting for you — not the other way around

Guardrail: Reads and summarizes — it never replies, deletes, or books without your approval.

Build this agent

Inbox Triage

Inbox zero, without the afternoon it used to take.

  • Sorts incoming mail: needs-you-now, can-wait, and noise
  • Drafts replies in your tone for the routine ones — you approve each send
  • Identifies unsubscribe candidates and lists them for your sign-off
  • Summarizes long threads so you catch up in seconds

Guardrail: Drafts and sorts — nothing is ever sent or deleted without your explicit approval.

Build this agent

Research Digest

Your topics, tracked — with receipts.

  • Tracks the topics, industries, or hobbies you specify
  • Sends a weekly digest of what actually changed — with linked sources
  • Distinguishes signal from rehashed noise and says why
  • Builds a running, searchable summary of everything it's found

Guardrail: Every claim links to its source — no unsourced summaries, no filler.

Build this agent

Life Admin

The appointments-and-paperwork assistant you never had.

  • Keeps a watchlist of appointments, renewals, and deadlines — and nags you early
  • Drafts the emails and form text for routine admin, ready for your send
  • Compares options (flights, insurance, subscriptions) in a plain table
  • Remembers your preferences so every draft starts 90% done

Guardrail: Prepares everything, commits nothing — bookings and purchases wait for your yes.

Build this agent
Browse all 12 templates

From zero to personal agent in three steps

01

Pick a template

Morning brief, inbox triage, research digest, life admin — or describe your own in plain English.

02

Get your agent's blueprint

The Builder generates the full agent — instructions, tools, model, and approval rules baked in.

03

Put it to work

Test it in the sandbox on the free trial, then run it on a schedule with approval gates on.

Cheaper than the hours it gives back

Start on the free trial — 1,000 credits, no card required — and every template above works on it. Bring your own API key (Anthropic, OpenAI, or Moonshot/Kimi) and it's always free with zero markup. Want your agents running 24/7 on a schedule? Paid plans start at $14.99/mo — with a hard spend cap you set.

See honest pricing

Questions people actually ask

Will it send emails or book things without asking me?

No. Approval gates are the default posture of every ME template: the agent drafts and prepares, and anything that sends, books, or spends waits for your explicit approval. You can review exactly what it wants to do before it does it.

How much does a personal agent cost to run?

Start free: the trial includes 1,000 credits (1 credit = $0.001 of real compute), enough to build and test your first agents. Bringing your own API key is always free with zero markup. If you want hosted, always-on agents that run on a schedule, paid plans start at $14.99/mo — and the exact cost is shown before every run.

Do I need to be technical?

No. Pick a template and the Builder turns it into a working agent — no code, no config files. If you can describe what you want in a sentence, you can build the agent.

Which apps can my agent connect to?

The connector catalog covers popular email, calendar, and productivity apps, and it's growing. Your agent can also work from text and links you give it directly — no connector required to get value on day one.

Is my data private?

With bring-your-own-key, your API key stays in your browser and you pay the model provider directly. Hosted runs keep an audit trail you can inspect, approval gates mean you see every sensitive action before it happens, and a kill switch stops everything instantly.

Which AI models power my agent?

Frontier models only — Claude Opus for blueprints, GPT-5.5 for sandbox chats, and bring-your-own-key support for Anthropic, OpenAI, or Moonshot (Kimi). No silent downgrades to cheap models, ever.

Get your evenings back.

Build the personal agent that handles the admin — while every decision that matters stays in your hands.

Build my personal agentTry Morning Brief first