Gaggo AI'yi adım adım öğren Learn Gaggo AI step by step

What is this?

Gaggo is your own AI software studio, running on your Mac. Not a single chat box — it's a team of 14 AI specialists plus a CEO that coordinates them. You say "I want an app like this," and the team turns it into a real, shippable app — planning the tasks, writing the code, reviewing it, and checking its security.

Two things make Gaggo different:

• It runs on your own subscription. It uses the Claude or Codex (ChatGPT) subscription you already pay for. No extra "credits" or per-use fees.
• Everything stays on your Mac. Your code, files, and data never leave your computer.

What happens on first launch?

The first time you open Gaggo, a short setup wizard welcomes you (5 steps). Press "Next" to move through each one:

1. Welcome — a quick summary of what Gaggo does.
2. How it works — meet your team and its departments.
3. Connection — choose which AI to use (Claude or Codex). In most cases you don't need to do anything; your desktop subscription is detected automatically. (You can enter your own key if you like — but it's optional.)
4. Add your first app — introduce the app you'll work on. Pick an existing folder, or just type your idea and start from scratch.
5. Budget & notifications — set a daily spending limit (so there are no surprise bills) and turn on notifications.

💡 Want to see this wizard again later? Settings → Setup wizard → Re-run the wizard.

Where do I land after setup?

When setup is done, you arrive at the Overview screen. From here:

• The left menu takes you to every section (Tasks, Activity, Company, Cost, Settings…).
• The bottom-left shows your remaining trial and your daily spend.
• The top lets you switch language (🇹🇷 🇬🇧 🇩🇪 🇪🇸).

What's next?

1. Haven't added an app yet? → Add your first app
2. Want to meet your team? → Your company — 14 AI agents
3. Ready to get work done? → Tasks (Board)

Tips

• Take it slow: try a small task first and watch how your team works.
• Start with a low budget limit (e.g. a few dollars a day); raise it as you get comfortable.
• Before doing anything risky (like deleting a file), the team asks for your approval. The final say is always yours.

What is this screen for?

This screen lets you introduce your first app to Gaggo so your AI team has something to work on. There are two ways in: connect an existing project folder on your computer, or start from just an idea and build everything from scratch.

How to open it

Click New App in the left menu. The window that opens shows two cards: Add from folder and Start from an idea. Pick one and you'll continue down that path.

Step by step: Add from an existing folder

1. Click the Add from folder card.
2. Type a name (e.g. "Supp Match"). The short ID box next to it fills in automatically — change it by hand if you like.
3. Optionally add a short tagline (what your app does in one line).
4. Type the platform (e.g. iOS · SwiftUI, Next.js, React Native…).
5. Press Pick folder and point it at the folder where your project lives.
6. Choose an emoji and an accent color — that's how you'll spot your app in the list.
7. Press Add. Your app is now on your board.

💡 The save button stays disabled until the folder path is a full path. Make sure you've picked the right folder.

Step by step: Start from an idea

No code yet? This path is for you. It walks you through four short steps and ends with your team getting to work.

1. Click the Start from an idea card.
2. Enter a name, a short ID, and the platform. Then write your idea in your own words — what it should do, who it's for, why it matters. (You'll need at least a few sentences; a character count sits under the box.)
3. Press Analyze. Your product manager Nova reads the idea and gives you an honest read: pros, cons, hidden assumptions, likely failure points, and a cheap way to validate it. At the top you'll see a clear verdict — Go / Iterate / Kill.
4. Optionally run a competitor analysis — the team pulls together similar products, their pricing, and gaps left open in the market. Want to skip it? Press Skip to confirm.
5. On the last step, pick the target folder (an empty folder is fine — the team sets up a working skeleton inside), review the summary, and press Create app.
6. That's it. The app is created and Nova immediately starts turning your idea into the first round of tasks.

💡 The clearer you describe your idea, the sharper Nova's verdict — and the first tasks it produces — will be.

Tips

• The short ID is generated for you as you type the name; only edit it by hand if you want a custom one.
• On the Start from an idea path the folder can be empty — the team scaffolds a working iOS skeleton for you.
• If Nova's verdict comes back Kill, don't panic; it's not burying your idea, it's surfacing the risks you couldn't see while you were close to it.
• Everything runs on your own computer and on your own Claude/Codex subscription — your idea never leaves.
• The color and emoji are purely cosmetic; you can change them anytime later.

What's next?

• Once your app is added, to keep a finger on the pulse of everything → Overview

What is this screen for?

Overview is the first thing you see when you open the panel — it's your daily snapshot. At a glance you can see how your apps are doing, what your team is working on, and what to focus on today.

How to open it

Click Overview in the left menu. When the panel opens, you're already here. At the top it greets you based on the time of day ("Good morning", "Good evening"…) and gives a short summary underneath: how many apps you have, how many people are on your team, how many tasks there are, and what percentage is done.

Step by step: Generate today's briefing

The big card at the top holds the CEO Briefing. This is a short plan from your product manager Nova that tells you what to do first today.

1. Press "Generate briefing".
2. Nova scans all your apps and produces a prioritized daily plan (about 30 seconds).
3. A summary line and 3 ranked recommendations appear on the card — the top one is the most urgent.
4. Want a fresh plan later? Just press "↻ New briefing" in the top right of the card.

💡 This runs on your own Claude/Codex subscription, on your own computer — everything stays with you.

Step by step: Add a new app

1. Press "New App" in the top right.
2. Pick a project folder on your computer (or start from an idea).
3. Your app appears in the card grid below, ready for your team to work on.

Reading the card grid

Under the Apps heading, each app has its own card. On every card you'll see:

• A health score next to the heart icon — your app's overall condition; hover it to see what's pulling the score down.
• Progress as a percentage and the task count (e.g. 6/20 tasks).
• How many tasks are active right now, if anything is running.
• A red blocker badge if there's an unresolved blocker.

The small icons in the top right of the card:

• CLAUDE.md export — writes the notes your team keeps for this app into the project folder as a file.
• Edit — change the app's name, icon and details.
• Remove — takes the app off the panel (it asks for confirmation first).

Reading the Daily Standup

Below, the Daily Standup splits into three columns:

• In progress — tasks being worked on right now, and which team member has them.
• Blocker — stuck work that needs sorting out. If this is empty, great.
• Up next · high priority — the most critical work waiting in line.

At the very bottom, the Recent runs list shows what your team did most recently.

Tips

• First thing in the morning: generate the briefing, then start with the top recommendation.
• If the Blocker column is full, look there first — that's what's slowing things down.
• If the health score dips, hover the heart icon to see why.
• The "Runs last 24h" number in the KPI strip is a quick read on how much your team worked today.

What's next?

• To meet your team → Your company — 14 AI agents

What is this screen for?

Company shows your AI team as an org chart. At the top there's you (the chair) and a CEO who coordinates everything; below them, specialists are spread across 5 departments, from engineering to design. This is where you get to know everyone, decide who does what, and tweak their settings.

How to open it

Click Company in the left menu. At the top you'll see how many people are on your team and how many are working right now.

Meet the team: hierarchy and departments

The top of the page shows the top of the org:

• Chair — that's you. You set the direction.
• CEO (Oliver Hayes) — your AI orchestrator. When you hand over a job, the CEO routes it to the right person.

Below them the team splits into 5 departments, each led by a director (the one with the crown badge):

• 💻 Engineering — writing code, review, and security. Ethan Brooks writes the code, Liam Carter reviews it, Marcus Hale handles security.
• 🧭 Product & Data — planning, metrics, and quality. Maya Sullivan is your product manager.
• 📣 Growth & Customer — store optimization, content, and customer relationships.
• 🎨 Design — interface and visual identity. Leo Marsh designs the UI.
• ⚙️ Operations — releases, translation, and finance.

Step by step: Get to know a teammate

1. Click anyone's card — a panel slides in from the right.
2. At the top you'll see their name, title, and a short intro.
3. Below that you'll see which tools they can use, which skills they have, and their recent work.
4. To close the panel, click outside it or hit the X in the corner.

Step by step: Hand someone a job

1. Press "Assign task" in the top right (or the "Assign task" button inside someone's panel).
2. Tell them what you want done in the message box that opens.
3. Send it. The job goes to the person you picked; watch its progress on the Activity screen.

💡 Not sure who to pick? Just hit "Assign task" and the CEO will route the job to the best-fit specialist.

Step by step: Change someone's settings

1. Click a card to open their panel.
2. Press the little pencil on the Model box — choose which brain they'll use: opus (most powerful), sonnet (balanced), or haiku (fast).
3. Use the permission mode box to set how freely they work: plan only, normal, or auto-accept changes.
4. In the Skills section, click the badges to switch extra expertise on or off for that person. You can always hit "Back to auto" to restore the original setup.

Step by step: Install a new skill

1. Press "Add skill" in the top right.
2. Pick one of the ready-made packs (for example the interface-quality or design-critique pack) and click a skill badge inside it.
3. You can also paste your own source to install. It takes a few seconds; after that, the skill is available to your team.

Step by step: Create your own agent

1. Press "New agent" in the top right and define a specialist from scratch.
2. Or go to someone's card and press the copy icon to spin off a new one based on them.
3. You can later edit or delete the ones you create (the built-in team members can't be deleted).

Tips

• The little dot on each card shows status: green means idle, amber means working. A working person's card also shows what they're doing right now.
• The crown director badge marks the head of each department — for big jobs, it's smart to start with them.
• Cards that look faded and can't be clicked are teammates that unlock with a higher plan.
• Your whole team runs on your own Claude/Codex subscription, right on your machine — your data stays with you.
• You don't always need the most powerful model: for routine work, haiku or sonnet is faster and cheaper.

What's next?

• To hand work to the team and run it → Tasks (Board)

What is this screen for?

The Board is where all the work for your app lives — like a sticky-note wall. Each task is a card, and cards move left to right: Planned → To do → In progress → Review → Done. This is where you create tasks, hand them to your team, and run them.

How to open it

Click Tasks in the left menu. If you have more than one app, use the tabs at the top to choose which app's board you're looking at.

If the board is empty — start with a Deep Scan

A new app starts with an empty board. Press the "Start deep scan" button in the middle:

1. Your AI team reviews your app from every angle (design, code, security, growth…).
2. Within a few minutes it fills the board with to-dos and suggestions.
3. Now you have a real list to work from.

Step by step: Add a new task

1. Press "New task" in the top right.
2. Write a title (e.g. "Simplify the sign-in screen").
3. Optionally add a description: what should happen, why, which screens are affected.
4. Pick a priority.
5. Press Create. The card appears in the "To do" column.

💡 Many tasks at once: in the description field you can write one task per line and create them in bulk.

Step by step: Run a task

1. Double-click a card — the detail window opens.
2. Choose the assigned agent (who should do it). Not sure? Leave it blank and the team auto-assigns.
3. Press Run. The task starts; watch its progress live on the Activity screen.

What do the top buttons do?

• Assign to team — auto-assigns unassigned tasks to the best-fit team member.
• Run in order — runs assigned tasks one by one in safe (plan-only) mode.
• Deep scan — the team reviews your app and produces new tasks and suggestions.
• Flow — starts a ready-made, multi-step workflow (see Workflows).
• Simplify — rewrites task titles in plain, easy-to-read language.

Tips

• Drag a card and drop it into another column to change its status.
• Break a big job into subtasks, and the team runs them in order through different specialists.
• For a risky task (one that changes files automatically), the team asks for your approval first — no surprises.

What's next?

• To see what happens while a task runs → Activity — watch & approve

What is this screen for?

Activity is where you watch what your team is doing live. When you run a task, it shows up here; you follow what they do step by step, see what changed, and give your approval when needed.

How to open it

Click Activity in the left menu. Runs are listed on the left; click one to see its details on the right.

How is the screen organized?

• Waiting for your approval — the team pauses here and asks before doing anything risky.
• Currently running — work that's active right now.
• Completed — finished work.
• Up next — planned work that hasn't started yet.

Step by step: Answer an approval

When the team reaches a risky step — deleting a file, installing a package — it pauses and asks:

1. Find the yellow notice in the Waiting for your approval section.
2. Read what it wants to do.
3. Press Allow (✓) or Deny (✕).

💡 If you don't answer, the request is auto-denied after a time limit you set (for safety). You can change this window in Settings.

Step by step: See a run's result

1. Click a run on the left.
2. On the right you'll see the steps (what the team did) and a summary.
3. If the team changed files, the "Repo changes" section shows what changed (the diff) — and you can save it with Commit if you like.
4. Details like duration and cost are here too.

Handy buttons

• Retry — restarts a failed run.
• Stop — stops a running job.
• Take snapshot — creates a restore point before changes.
• Diagnose — if a job failed, Vera (the code reviewer) explains the cause in plain language (~10s).
• Promote to Atlas — turns a suggestion-only step into real code, written by Atlas.

Tips

• The team often fixes itself: if some code won't build, it spots the error and retries on its own.
• If you can't tell why a job finished or failed, press Diagnose — no need to read technical logs.
• Before an important change, Take snapshot; if it goes wrong, you're one click from rolling back.

What is this screen for?

Quick Start is the cockpit where you turn an idea into a working MVP. Your AI team turns the design, the screens and the data layer into real code — not just suggestions, but real files. You run a left-to-right build plan with a single click.

How to open it

Click Quick Start in the left menu. If you have more than one app, use the tabs at the top to choose which one you're building. If you don't have any app yet, you'll be asked to add an idea first.

What does the top card show?

The card at the very top is the readiness of the selected app:

• The ring on the left gives a health score out of 100 — how ready the project is to ship.
• Next to the ring are iOS controls: small buttons that let you build your project, test it, or open it straight in the simulator. (If there's no connected iOS project, these buttons are dimmed.)
• Below, a progress bar shows how many tasks are done as a percentage.
• If there are risky spots, they show up here as small tags.

Step by step: Run the Build Plan

The plan has three columns, each with a number:

1. 1 · Interface — design identity, MVP screens, localization.
2. 2 · Data & Service — outside data source, backend, wiring screens to data, paywall.
3. 3 · Security & Release — security scan, testing, ASO, analytics, launch prep.

Here's how you move through it:

1. Start from the top step of column 1 — press Run next to it.
2. A step won't open until the one above it is done. Locked steps are greyed out; they unlock when it's their turn.
3. Once a column is finished, move to the next number: 1 → 2 → 3.
4. Only one step runs at a time. While something runs you'll see a "a step is running" badge at the top and the other buttons lock briefly — this prevents conflicts.

💡 The first step is always Design identity: Atlas turns the color palette, theme and Welcome/Onboarding screens into real code. Start here for a strong foundation.

Step by step: Turn MVP screens into code

1. First add an idea to your app — Nova, your product manager, drafts the screen plan.
2. In the Interface column, the MVP screens step shows how many screens are ready with a small counter (e.g. 3/8).
3. Press Run — Atlas codes all your MVP screens in priority order, without conflicts.
4. When it's done the step turns green; everything runs on your own machine, on your own Claude or Codex subscription.

Step by step: Build and test the iOS project

1. On the readiness card up top, use the small buttons next to the ring.
2. The shield icon builds your project, the tube icon runs the tests, and the play icon opens the app in the simulator (builds + installs + launches).
3. If there's no connected iOS project, the caption reads "no iOS" and the buttons are dimmed.

Tips

• Work the plan in order: 1 → 2 → 3. Finish each column top to bottom instead of skipping around.
• You can rerun a step — after it's done the button becomes "Rerun".
• The MVP / Blocker / Last bar up top gives you a quick pulse: how many screens, any open blockers, what ran last.
• If you click a locked step, you'll see a tip explaining why (finish the one above / requires a higher plan).
• At the bottom, Risk Signals warns you about blockers and errors that could trip you up before launch.

What's next?

• To start each morning with a ready-made digest → Briefing — your daily digest

What is this screen for?

The Briefing is a daily digest written for you by Nova, your product manager. She scans all your apps one by one, ranks the work, and tells you in plain language what's worth doing today. You sit in the boss's chair and read a single report.

How to open it

Click Briefing in the left menu. The first time you open it the screen is empty — you need to generate a briefing.

Step by step: Generate your first briefing

1. Press the "First briefing" button in the top right (later it becomes "New briefing").
2. Nova reviews all your apps in about 30 seconds.
3. You get a single report: a headline at the top (your overall status in one sentence), then suggestions and app statuses below.

💡 The briefing runs entirely on your own Claude/Codex subscription; each one counts toward your selected AI provider's usage. Everything stays on your own machine.

Step by step: Read the briefing

The report is laid out to be read top to bottom:

1. Headline — the day in one sentence. In a hurry? Read just this.
2. Suggested order — the work Nova has ranked, as 1, 2, 3… Each suggestion shows why it matters and which app it's about.
3. Cross-app insights — things that touch more than one app at once and deserve attention (the yellow warning cards).
4. App status — for each app, a short status, a momentum tag (high momentum, steady, low pace or blocked), any blockers, and the next step.
5. This period — a bullet list summing up what happened over the period.

Step by step: Chat with Nova

If something in the briefing isn't clear, just ask:

1. Scroll to the "Chat with CEO" section under the report.
2. Type into the box — for example "explain this in more detail", "why is this a priority?", or "what should I do for that app?".
3. Press Send (or ⌘/Ctrl + Enter on the keyboard). Nova answers with the context of that same briefing.

Step by step: See previous briefings

Generating a new briefing doesn't wipe the old ones. The "Previous briefings" list at the bottom of the report shows the headlines and dates of past digests, so you can track your progress over time.

Tips

• 💡 Generate a briefing first thing in the morning; you'll have your plan for the day in 30 seconds.
• Don't puzzle over a suggestion you don't get — ask Nova. She remembers the full context of that briefing.
• Give priority to any app tagged blocked; that means something is stopping progress there.
• In the rare case the headline comes back empty, check the "This period" block or run the briefing again.

What's next?

• For a longer-term view of the work Nova suggests → Roadmap

What is this screen for?

The Roadmap is a timeline that shows where your app is heading, version by version. Each stop (a milestone) is a release — like v1.0 or v1.2 — and lists what goes into it. You don't fill anything in here; the stops move forward on their own as your team finishes the work.

How to open it

Click Roadmap in the left menu. If you have more than one app, use the tabs at the top to choose which app's roadmap you're looking at.

Step by step: Read the roadmap

1. Pick your app at the top. The timeline shows only that app's stops.
2. Stops are ordered top to bottom, from earlier versions to later ones.
3. Each stop shows: a version tag (e.g. v1.2), a title, a short goal sentence, and the work items for that release.
4. The colored dot on the left tells you that stop's status.

Step by step: Understand the statuses

Each stop has a small label on the right and a colored dot on the left:

1. Done (green) — this version is finished; the team wrapped it up.
2. Now (amber, gently pulsing) — the version being worked on right now.
3. Planned (gray) — a future version, not started yet. It stays a quiet gray so it doesn't feel like it's already in progress.

💡 The version tags (v1.0, v1.2) sort the timeline automatically — so the "Now" stop always sits just below the finished ones and just above the planned ones.

If the roadmap is empty — start with a Deep Scan

A new app starts with an empty roadmap. The screen suggests running a Deep Scan. You kick off a Deep Scan from the Tasks screen:

1. Your AI team reviews your app end to end.
2. It proposes a sensible version plan for you — which work belongs in which release.
3. When you come back to the Roadmap screen, the stops are filled in.

Tips

• This is a read screen; you don't click around to change things here. You do the real work on the Tasks screen, and the roadmap reflects it.
• As you finish tasks, stops promote on their own — "Planned" turns into "Now," then into "Done" over time.
• A stop's goal sentence reminds you why that version exists at a glance — so you keep the big picture even in long lists.
• If you have several apps, hop between the tabs to quickly compare where each one stands.
• Everything runs on your own computer and uses your own Claude/Codex subscription — your roadmap never leaves your machine.

What's next?

• To see what your team remembers and why → Memory

What is this screen for?

Memory is the shared notebook your AI team reads before starting any job. The info, rules, and lessons you write here (e.g. "prices should come dynamically", "keep this screen simple") are handed to the team on every run — so you never have to repeat yourself.

How to open it

Click Memory in the left menu. At the top you'll see how many notes you have, how many are pinned, and a reminder that the team reads them on every run.

Step by step: Add a new note

1. In the box at the top, first pick a scope: General (whole team) or one of your apps.
2. In the field next to it, write a one-sentence info or lesson (e.g. "Paywall prices should come dynamically from RevenueCat").
3. Press Add (or hit Enter). The note drops under the scope you chose.

💡 General notes are read by everyone; app-specific notes only kick in when the team is working on that app.

Step by step: Pin a note

1. Hover over a note and press the pin icon on the right.
2. The pinned note moves to the top and is kept safe during deletions.
3. To undo, press the same icon again to unpin.

💡 Pin your most critical rules — that way bulk clean-ups never wipe them by accident.

Step by step: Edit and delete a note

1. Hover over a note; a pencil (edit) and a trash (delete) icon appear on the right.
2. Press the pencil, change the text, then press the check mark to save (or the X to cancel).
3. Press the trash icon to remove that note.

Step by step: Search and filter notes

1. Type a word into the search box at the top — notes filter instantly.
2. Use the All / Manual / Agent buttons to filter by source: did you write it, or did the team add it.
3. Use the Pinned button to show only pinned notes.
4. When you're done, press Reset to clear all filters.

Step by step: Bulk delete

1. Press the Select button in the filter bar — this enters select mode.
2. Tick the checkboxes next to the notes you want to remove.
3. Press Delete selected and confirm. Pinned notes are still kept.

What is the Memory Map?

As your notes grow, a small Memory Map appears at the top. It shows at a glance how your memory spreads across apps, where notes came from, and which topics are trending — and the team puts these signals to good use too.

Tips

• Keep it short and clear: one-sentence, definite rules are the language the team understands best.
• When agent notes pile up for an app, a "clear intake notes" button shows up in that section; tidy up now and then to keep things clean.
• Pinned notes survive every deletion — don't forget to pin your most important rules.
• Everything runs on your own computer and your own Claude/Codex subscription; your notes never leave.

What's next?

• To see how much your team is spending and where you can save → Cost & savings

What is this screen for?

This screen shows how much your AI team has been working and what it costs you. Best of all: it compares your subscription against your real usage and shows, in plain numbers, how much you're saving.

How to open it

Click Stats (or Cost) in the left menu. Use the 7d / 30d / 90d / All buttons in the top right to pick the time range — the whole screen updates to match.

Step by step: Read the period summary

The four boxes at the top give you the headline numbers for the range you picked:

1. Period cost — total spent in this range.
2. Runs — how many tasks your team ran (and the average cost per run).
3. Month projection — where you'll end up by month-end at this pace.
4. Wasted — money spent on runs that ended in errors. If the share climbs, the box turns red — a sign something is getting stuck.

Step by step: Track today's budget

If you set a daily spending limit, you'll see the Today bar:

1. The bar shows how much of today you've used ($ used / $ limit).
2. It turns yellow past halfway and red as you near the limit.
3. The line below tells you how many runs happened today and the average cost.

💡 No bar showing? You haven't set a daily limit yet — you can add one from Settings.

Step by step: See your savings

The big card in the middle is the fun part. Gaggo runs on your own Claude / Codex subscription, and this card works out what the same work would have cost if you paid for each piece separately:

1. Subscription — what your subscription costs you over the selected range.
2. Equivalent usage — what the same work would have cost piece by piece.
3. Profit / shortfall — the gap between the two, i.e. what you kept.
4. Break-even bar — fills as you earn back your subscription; turns green once you pass it.

If your savings are positive, the card shows a "you got this many times your money back" headline. The Share button at the bottom right lets you copy and share those numbers.

💡 No runs yet? The card shows an estimated preview with an EXAMPLE badge. After your first real run, it switches to your true numbers.

Step by step: Read your team's performance

The Team performance section shows who's pulling their weight:

1. The three boxes up top: success rate, total errors, and crew cost.
2. The list below ranks each team member (Nova, Atlas, Vera, Sentinel…) with their own success percentage, new work in the last 7 days, average duration, and cost.
3. A green bar means runs are going smoothly; a yellow bar means that member hit errors.

Step by step: Dig into the details

Further down the page are cards that break the work apart:

1. Cost trend — a chart of your daily spend; the peak is noted at the top.
2. Cost by agent — how much each member spent.
3. By app — which of your apps drew the most runs.
4. Most expensive runs — your range's 10 priciest tasks, one by one.
5. Daily table — recent days' run counts and costs, day by day.

Tips

• If a range looks empty, don't worry — it just means no runs in that window; pick All up top to see your full history.
• If Wasted stays high, look at the failing tasks; most of the time a small tweak fixes it.
• Everything runs on your own machine and your own subscription — these are your real numbers, shared with no one.
• The Most expensive runs list is the fastest way to spot the tasks eating the most budget.
• As your savings grow, the card may invite you to a higher plan; if you'd rather not, just ignore it.

What's next?

• To get a heads-up when a run finishes or needs your approval → Notifications

What is this screen for?

Notifications is where all the news from your AI team collects — approval requests, run errors, finished briefings, and other important events, all in one list. Check in here now and then to make sure nothing slipped past you.

How to open it

Click Notifications in the left menu. The header shows the total number of records and how many are still unread.

What's in a notification?

Each notification shows up as a card, and its color tone tells you how important it is:

• Red tone — something critical (e.g. a run error or a security-related abort). Sentinel, your security guard, surfaces events like these.
• Yellow tone — a warning worth your attention (e.g. an approval timeout or a budget limit).
• Plain tone — just for information (e.g. a new briefing is ready).

Each card also has small labels: the kind of event (like Approval needed, Run error, New briefing) and, when relevant, which app it belongs to. Unread cards have a little dot glowing next to them.

Step by step: Focus on one kind of notification

1. Look at the filter labels above the list — each one shows how many records of that kind exist.
2. Press All to see every notification.
3. Press Unread to keep only the ones you haven't seen yet.
4. Press a kind label like Approval needed or Run error to narrow the list to that kind.

💡 Only the kinds that actually exist right now appear as labels — if there are no run errors in the list, you won't see that label.

Step by step: Mark a notification as read

1. Press "Mark read" in the bottom-right corner of the card.
2. The card fades and the unread count drops by one.
3. Want to clear them all at once? Press "Mark all read" in the top right — one tap marks the whole list as read.

Step by step: Clear the list entirely

1. Press "Clear" in the top right.
2. Confirm when the prompt asks.
3. All notifications are deleted and the list empties. It fills back up as new events happen.

Tips

• If you see an Approval needed notification, you give the actual approval on the Activity screen; the card here is just a reminder.
• Read the red-tone cards first — they usually point to work that's stuck or was stopped for safety.
• New briefing notifications tell you a summary your team prepared for you is ready.
• Everything runs on your own computer, on your own Claude/Codex subscription — notifications never leave your machine; they're just for you.

What's next?

• To automate repeating, multi-step jobs → Workflows & templates

What is this screen for?

Workflows let you kick off a multi-step job in one click. You pick a ready-made template, type a topic, and your AI team hands the work down the line — for example, Atlas writes the code first, Vera reviews it, then Sentinel checks the security.

How to open it

Press "Flow" in the top buttons on the Board (see Tasks). You can also open it from the Overview screen or the quick command box (⌘K). To build your own templates, go to the "Workflow Templates" card in Settings.

Step by step: Start a ready-made flow

1. Press "Flow" — the template list opens. Each template shows a short description and a "when to use" note, with the team members that will run in order listed underneath.
2. Click the template you want to run.
3. In the topic box that appears, type what the work is about (e.g. "checkout screen").
4. Press Start. A parent task is created with step-by-step subtasks underneath it, and the team runs the whole chain from start to finish. Watch progress on the Activity screen.

💡 Pressing Enter in the topic box also starts the flow — quick and one-handed.

Step by step: Tune a flow before running it

After you pick a template, you don't have to use every step exactly as-is:

1. Use the person list next to a step to hand that job to a different team member.
2. To leave a step out, press Skip — it turns grey and won't run in that flow. The header shows how many steps are skipped and how many are still active.
3. To change a step's instructions, press the pencil icon and edit the title and description text. Anything you write as {topic} is swapped in for your topic automatically.
4. Changed your mind? "Reset to default" puts that step back the way it was.

💡 When you make changes, a "Save as template" option appears at the bottom — it stores your tuned flow as your own template so you never have to set it up again.

Step by step: Build your own template

1. Open Settings → Workflow Templates. The 5 ready-made templates sit there with a lock badge (built-in); these can't be edited, but you can duplicate one with the copy icon and tweak it to fit you.
2. To start fresh, press "New template".
3. Give it an emoji, a name, a short description, and a when to use note.
4. Use Add step to lay out the steps: which team member runs it, and its title and instructions. Move steps up or down with the arrows, or remove one with the trash icon.
5. A live preview chain appears at the bottom. Press Create; the template joins your "Flow" list.

💡 The pencil and trash icons only show on the custom templates you create — the built-in ones stay safe.

Step by step: Get and share gallery templates

1. In the Workflow Templates card, press "Gallery". Ready-made flows shared by the community are listed.
2. On a template you like, press "Import" — one click adds it to your own templates.
3. To share your own template, use the "Share" link at the top.
4. To pass a template to someone, the download icon on its row copies it; the other person loads it back with "Import".

Tips

• Templates run on your own Claude/Codex subscription and on your own computer — everything stays local.
• On a risky step, the team asks for your approval first before changing any files — no surprises.
• To keep a flow lean, Skip the steps you don't need, then "Save as template".
• If you leave a field blank while importing, it's safely filled in with sensible defaults.
• Your custom templates are a Studio+ feature; if it's locked, an upgrade window opens.

What's next?

• To build your subscription and payment screen → Paywall Builder

What is this screen for?

Paywall Builder is where you design your app's purchase screen (paywall). You pick a feel, packages, and perks; as you fill in the form on the left, a real phone preview updates instantly on the right. When you hit "Create," your AI team writes the real SwiftUI + RevenueCat code from that design.

💡 This feature is only available on the Founder plan.

How to open it

Open Tasks from the left menu. Click the Paywall button in the top bar — the design window opens. (If you're not on the Founder plan, you'll see a lock 🔒 next to the button.)

Step by step: Design the paywall

1. In the top "Describe the paywall" box, write what you want in your own words (e.g. "7-day free trial, then highlight the yearly plan").
2. Pick a feel: Trial offer, Bold, Minimal, Premium, Hard gate, Playful, Value stack, or Social proof. Each feel changes the look and mood of the screen.
3. Choose the free trial length: None, 3, 7, or 14 days.
4. Set up your packages (see the steps below).
5. In the Premium perks box, write one item per line (e.g. "Unlimited usage", "Ad-free experience").
6. Optionally open the Advanced section to add a headline, subheadline, button text, social proof, and an icon emoji.

💡 The Live preview on the right refreshes on its own with every change — you see the result as you type.

Step by step: Edit packages

1. Press Add package to create a new plan.
2. For each plan, fill in a label (e.g. "Yearly"), a duration (Weekly / Monthly / Yearly / Lifetime), and a price.
3. Optionally add a per-period breakdown on the second line (e.g. "$4.99/mo") and a badge (e.g. "BEST VALUE").
4. To feature a plan, tap the star — that package gets highlighted in the preview.
5. Remove a package you don't want with the trash icon.

💡 The prices here are for preview only. Real prices come from RevenueCat inside your app.

Step by step: Preview and generate the code

1. To see the design bigger, press Enlarge above the preview — the paywall opens in a new phone-sized window.
2. When you're ready, press Create at the bottom.
3. Your coder Atlas starts writing the real SwiftUI + RevenueCat code from the design, and the large preview window opens.
4. Watch the job's progress live on the Activity screen.

Tips

• To generate code, an app must be selected. With no app, you can still design and use Enlarge to see the phone preview.
• "Create" won't run until you have at least one package added.
• Writing your intent in the "Describe" box helps the team hit the right tone.
• Everything runs on your own Claude/Codex subscription and on your own computer — both the design and the code stay with you.
• Pick a feel and hit Enlarge to quickly try out a few different moods.

What's next?

• To manage plans, locks, and other preferences → Settings

What is this screen for?

Settings is the one place where you tune how Gaggo works — your plan, language, daily budget, safety rules, automations, and backups all live here. Most settings save automatically, so there's no separate "Save" button to hunt for.

How to open it

Click Settings in the left menu. The page is split top to bottom into sections: Plan & provider, Execution & safety, Automation & backup, Application, and Danger zone at the very bottom.

Step by step: Change the language

1. Look at the flag buttons in the top-right corner.
2. Click the language you want.
3. The interface switches instantly — and your AI team will write to you in that language from now on.

Step by step: Activate your license

1. Go to the License card. The strip at the top shows how many trial days you have left, or that your license is active.
2. Once you buy a plan, the provider emails you a license key.
3. Paste that key into the License key box. You can add your email too (optional).
4. Press Verify and activate. Your license goes active and all your plan's features unlock.

💡 Moving to another machine? Use "Remove from this device" to release your license.

Step by step: Pick your AI provider

1. The AI provider card has three options: Auto, Claude, and Codex.
2. Auto is recommended — it finds your ready session automatically (Codex/ChatGPT if connected, otherwise Claude).
3. Choose Claude to have your team run on your Claude subscription.
4. Codex is only selectable if you have a Codex session signed in on your computer.
5. In Auto mode you don't need to enter any key — your ready desktop session is used.

💡 Just below, in Your AI subscription, mark which plan you pay for (e.g. Claude Pro). The Cost screen uses that number to show how much value you're squeezing out of your subscription.

Step by step: Set a daily budget

1. In the Budget card, pick one of the preset amounts ($5, $10, $25, $50) or choose Unlimited.
2. To type your own figure, use the box next to the presets.
3. If you set a limit, you'll see how much was spent today on the live bar.

Step by step: Permission mode and approval timeout

1. The Default permission card has three modes: Plan (only drafts a plan, touches no files), Default (asks for your approval on risky steps), and Accept edits (applies edits automatically).
2. If you like to play it safe, stay on Plan or Default.
3. In the Approval timeout card, set how long the team waits for you when it needs an approval (e.g. 30 minutes).

Step by step: Turn on automations

1. The Automations card has three ready routines: - Daily brief — prepares a short summary every morning at the time you set. - Weekly intake — gathers fresh work ideas on the day and time you choose. - Stale blocker check — catches work that's been waiting longer than the number of days you set.
2. Tick the one you want and adjust its time/day/number field. The rest happens on its own.

💡 Automations unlock with the Studio plan. If locked, the badge on the card tells you which plan you need.

Step by step: Back up and restore

1. In the Auto snapshot card, pick Off, Daily, or Per run.
2. To take a backup right now, press the Create button in the Backups card.
3. The list shows each backup's date and size. To bring one back, press Restore next to it; to delete one, click the trash icon (restoring asks you to confirm first).

Tips

• When you change the language, Nova (your product manager), Atlas (the coder) and the rest of the team start speaking it too.
• If you're new, keep the permission mode on Plan — the team changes nothing automatically and only drafts suggestions.
• Before an important change, take a manual backup; if something goes sideways you're one click from rolling back.
• Use Re-run onboarding to replay the intro tour whenever you like.
• "Wipe my data" in the Danger zone at the bottom can't be undone — since everything lives on your own computer, this clears all of it.

What's next?

• For details on plans, your trial, and licensing → Plans, trial & license

What is this screen for?

This is where you see which plan you're on, track your trial, and enter and activate the license key for the plan you bought. It all lives in the Plan & provider section at the top of Settings.

How to open it

Click Settings in the left menu. At the top you'll see three cards: the plans card (your current plan + upgrades), License, and Partner program.

Plans at a glance

Gaggo comes with three plans — there's no free tier. Indie is the entry door and includes a 3-day full-feature trial (card required); Studio and Founder are direct purchases.

• Indie — $19/mo (or $189/year, ~2 months free). Entry plan: 1 active app, 5 core agents (Atlas/Nova/Luna/Pixel/Devin), local database + live feed. All features unlocked for the first 3 days (card required, cancel anytime). After the trial it returns to the Indie limits.
• Studio — $39/mo (or $389/year, ~2 months free). Production plan: 3 active apps, the full 14-agent team, one-click MVP, deep scan, code review (Vera), security scan (Sentinel), competitor analysis, custom workflows, automations, auto-snapshots, run in simulator.
• Founder — $49/mo (or $489/year, ~2 months free). Studio plan: unlimited apps + everything in Studio + Paywall Builder (RevenueCat) + custom agent design + cross-app context + advanced analytics + visual feedback loop.

💡 Everything runs on your own computer with your own Claude/Codex subscription — the plan unlocks what Gaggo can do, not your AI bill.

Compare plans

Step by step: Start the 3-day Indie trial

1. On the plans card, pick a cycle with the Monthly / Annual toggle up top (annual has a −17% badge — ~2 months free).
2. Press the green button under Indie (e.g. "Upgrade to Indie").
3. The window that opens takes you to checkout. It asks for a card but you're not charged for 3 days; during this window you have full Studio + Founder access.
4. You can cancel anytime — cancel before the last day of the trial and you pay nothing.

💡 Studio and Founder are direct purchases — they don't offer trials. If you want to try first, start with Indie, then upgrade if you like it.

💡 When you tap a locked feature (like the Paywall Builder), a small upgrade window pops up too. The same button sends you to checkout, or use the "I have a license key" link to jump straight back here.

Step by step: Enter your license key

After buying, you'll get your key from the provider by email. To activate it:

1. Go to the License card. The colored strip at the top shows your status — Trial, License active, or Expired.
2. Paste your key into the license key box.
3. Optionally add your email next to it (not required).
4. Press Verify and activate. If the key is valid, the strip turns green and your plan shows on the card.

Step by step: Remove the license from this device

1. On the License card, when active, press the Remove from this device link at the right of the status strip.
2. The license is released from this computer; you can use the same key on another device.

Tips

• The trial strip shows how many days are left — decide before the last day.
• When the trial ends and you fall back to Indie, premium features (Paywall Builder, custom agents, etc.) lock; you can upgrade to Studio or Founder from the same card.
• After pasting the key you can just press Enter instead of clicking the button.
• Enter a wrong or expired key and a red warning shows under the box; copy the correct key from your provider email.
• If you're on Founder, the card celebrates it in gold and shows no upgrade button — everything is already unlocked.
• If you love Gaggo, the Partner program card lets you share your own invite link and earn a commission on every sale.

What's next?

• To see your AI spend → Cost — budget & spend
• Back to index