Everything it does, and everything it does not.

Install

Agentic PM is a desktop app. There is no terminal step and no developer tooling to buy.

The app installs as AI Second Brain. That is the desktop app behind Agentic PM, so that name in your Start menu or Applications folder is the right one.

Windows: the installer is not code-signed yet, so SmartScreen shows "Windows protected your PC". Click More info, then Run anyway.

macOS: the build is notarized by Apple. The developer shown is Juanita Vyatri, the developer account the app is published under.

It needs an AI engine: sign in with your Claude Pro or Max plan, or use the built-in free option. Keys stay on your machine.

The fourteen day trial starts at checkout on the pricing page. A card is required, nothing is charged today, and the first charge is on day fifteen. Cancel before then from Creem's receipt email and you are never charged. Sign in to the app with the same email you paid with.

Your first run

Point it at the meeting notes you already have. Nothing else is required.

1. Open the app and create a workspace. Name it after the product or team, not yourself.
2. Put your exported meeting notes in the workspace, or connect a note taker below.
3. Run /follow-ups. It lists the open loops and promises it finds.
4. Open the Work view. It shows what you owe, what you are waiting on and the decisions to make.

Connect a note taker

Agentic PM records meetings itself, on your computer, and no bot joins the call. The phone app can record too. It also reads what your existing note taker produces.

How it works: five stages, and every one reports its own health

Most AI tools are a blank chat box. You paste a task, you carry the answer back by hand, and nothing about your product is remembered. Agentic PM is a pipeline instead: work enters on the left, passes through four stages, and comes out as an action that waits for you.

The map below is an example of a fully set-up workspace. Every scheduled item in it is a routine you switch on yourself: out of the box, routines are off, and they run only while the app is open. The idea worth keeping is the last column: every part should say whether it is OK, idle or broken.

Senses: what comes in without being asked

Chat mentions, mail, calendar, meeting recordings, tracker state. None of this is something you fetch. It arrives, and the arrival is what starts everything downstream.

Reflexes: the parts that need no model

Sweeps, extraction, pre-meeting cards. These run on a clock and are deliberately mechanical, because the boring, structured majority of the week should not cost a model call or wait for your attention.

Brain: where judgment actually happens

The morning brief, the evening recap, meeting minutes, the weekly report. Model sessions are reserved for work that genuinely needs reasoning, which is what keeps this affordable on your own subscription rather than requiring a seat you buy from us.

Memory: what it knows between runs

Commitments, what you are waiting on, decisions, people, portfolio, and the context file that you extend every time you correct it. This is the layer that keeps a chat from starting from zero.

Hands: nothing moves without you

Posting, documents, calendar, tickets, direct messages. Every node here is permanently gated. Not "gated by default" as a setting you could switch off. The approval is structural, which is why the state is its own colour on the map.

Where the commitment ledger sits

The ledger is a file that your chats write. The app reads it and shows it in the Work view, and never writes it. Nothing checks it in the background: when you run /follow-ups, it reads your recent meetings and threads, lists the open loops and promises, and drafts the chase. Nothing is sent without your approval.

The ledger

The ledger is a file of your commitments that your chats write. The app reads it and never writes it. The Work view shows it as three lists:

• You owe. What you promised other people.
• Waiting on others. What other people promised you.
• Decisions to make. What is still undecided.

From any item you can start a chat to draft the chase. Nothing is sent without your approval.

The context file

One plain markdown file per workspace, CLAUDE.md, holding your rules: your product, your people, your vocabulary and the decisions already made. Every chat reads it before it starts.

You edit the file directly at any time, or run /remember to save a preference, a correction or a fact worth keeping.

How it finds what slipped

It does not watch in the background. Two things do the work, and you start both:

• The Work view lists what you owe, what you are waiting on and the decisions to make, read from the ledger.
• /follow-ups reads your recent meetings and threads and lists the open loops and promises it finds, then drafts the chase for your approval.

It is an AI model reading your notes, so check the list before you chase anyone. The four starter routines are off by default and run only while the app is open.

The eight jobs

Each job is a command or a feature you run yourself, or a routine you switch on. Routines are off by default and run only while the app is open. Every job in this table is in the build you can install today, with one exception that is marked as such.

The commands that ship with it

Skills are the things you invoke by name. Twenty six arrive in every new workspace, and any of them can be put on a schedule from the Routines tab without writing a config file.

You can add your own. A command is a markdown file in the workspace, so writing one is writing a document, and a command you delete stays deleted since 0.13.9.

Skills

Twenty six commands ship in every new workspace, in the groups above, included on every paid plan. They read the same rules file the jobs do, so output is about your product rather than a template. Every one is a plain markdown file you can open, change or delete.

Custom skills are plain markdown files you drop into the workspace. Yours are never uploaded.

More skills from the same author

These work with the same workspace. Since desktop 2.0.1, gdoc-surgical, ai-prototype-kit, artifact-comments, animated-readme and voiceprint ship in every new workspace. explainer-video is installed separately. Most are free and open source (Apache-2.0).

Export and portability

Everything exports to plain markdown and JSON, on every tier including free. There is no export paywall and no proprietary format.

If you cancel, the ledger stays on your machine and stays readable. Cancelling removes the agents, not your history.

Privacy and data

• The ledger is local. It lives on your device, not on a server.
• Recordings and notes stay on your computer. The exception is the text the AI reads, for example a transcript when it writes minutes. That goes to the AI provider you picked.
• Nothing trains anything. Your content is not used to improve any model.
• Team plans share context, not transcripts. The shared layer is the context file and the commitment ledger, both of which you can read and edit.

Known limits

1. It needs an AI engine to read your notes and write drafts: a Claude Pro or Max plan, or the built-in free option, which is slower and less exact.
2. No time-saving figure is claimed anywhere, because none has been measured. When one exists, taken on real work with a readable method, it will be published with the method.
3. The evidence so far is one operator's own meetings. Fifteen meetings, one team. Enough to prove the failure is real, not enough to say how often it hits you.
4. Nothing chases anyone for you. It drafts the nudge and you send it. There is no background watching: things run when you run a command or switch on a routine.
5. It does not put a bot in your calls. It records on your own computer, and the phone app can record too.

FAQ

Changelog