CourseGuidesArticlesPodcastGitHubAboutStart free course
AI Project Ideas: 30+ Things Real People Build/Work & office

Guided build · ~15 min

Reply drafts in your voice

Common email replies written in your tone, ready to tweak and send.

You will build this by describing what you want in plain words, in your own language. Claude writes the code; your job is to describe the result well and check it. Here is the whole path, start to finish.

ℹWhat you'll have at the end

Common email replies written in your tone, ready to tweak and send.

First, how to actually talk to the agent

  1. 1Describe the RESULT, not the tech: what's on screen, what each button does, what gets saved. The agent handles the 'how'.
  2. 2Be concrete about screens and fields — 'a form with amount, category and date' beats 'let me add expenses'.
  3. 3Ask for the smallest working version first. You grow it by talking afterwards — that's Step 2, not the first message.
  4. 4Decide what's YOURS, not the agent's: name the categories, colours, rules. Don't make it guess what only you know.

Step 1 — Build it live

Prompting for this kind of tool

  1. 1Nail THREE things: how templates are grouped (a sidebar by situation — refund, scheduling, follow-up), what happens when you click one (it opens in the main panel), and that placeholders like {name}, {date}, {company} turn into little fill-in fields at the top.
  2. 2Tell Claude to write the templates in YOUR tone — paste two or three replies you've actually sent so it copies your real voice instead of stiff corporate English.
  3. 3Decide yourself that the 'copy reply' button puts the FINISHED text (placeholders filled) on your clipboard in one click, and that everything saves on your device so your templates survive a refresh.
  4. 4End with: 'build the whole thing, check that it works, and tell me exactly how to open it.'
Start Claude
claude

Open the terminal and run this (if Claude is already running, type /exit first). Then ask it to make a folder for this project and work inside it.

The wrong ask vs. the right ask
✕ what most people type
you ▸ answer my emails
✓ what actually works — copy it, or better, say it in YOUR words
you ▸ Build me a reply-templates tool. A sidebar groups my saved email replies by situation (refund, scheduling, follow-up); clicking one opens it in the main panel, where any placeholders like {name}, {date} and {company} become little fill-in fields at the top. As I fill them the preview updates, and a 'copy reply' button puts the finished text on my clipboard in one click. I can add, edit and delete templates. Everything saves on my device. Clean, calm look, written in my own tone. Build the whole thing, check that it works, and tell me exactly how to open it.
✦✦

Say it in YOUR language — Russian, Spanish, anything. Claude understands.

Build it now — tick as you go

0/6

Step 2 — It's built. Now make it yours

ℹClaude said 'open localhost…' — now what?

Don't panic at the techy link. Claude tells you one of two things. Either a file like index.html — find it in your project folder and double-click it; it opens in your browser. Or a web address starting with http://localhost — copy it into your browser's address bar and press Enter. 'localhost' just means 'this is running on your own computer right now'. That's your tool. It's real.

Now tweak it until YOU like it — just talk

This is the part that makes it YOURS. Don't accept the first version — change it by saying what you want, right there in the same chat. Claude rebuilds and you refresh. Keep going until it fits how you actually work. For this tool, try things like:

you ▸ add a friendlier and a stricter version of each reply
you ▸ remember the last few people I wrote to
you ▸ add a {company} placeholder I can fill in
✦How I (Evgenii) actually do this

I don't type long messages. I open voice-to-text (the mic in ChatGPT, or your phone keyboard), and just TALK through everything I think about what it built — what's wrong, what I want next — then select-all, copy, and paste the whole brain-dump into Claude Code. Messy spoken thoughts work fine; the agent sorts them out. It's several times faster than writing.

Step 2.5 — Test it like a real thing (it won't be perfect first try)

The truth no one tells beginners: the first build is a DRAFT. The agent makes something, you run it, you find what's off, you tell it — and THAT loop is the actual work. Don't read the code. Use the tool like a stranger would and watch where it lets you down. For this tool, check exactly this:

  1. 1Open it (Claude gives a localhost link or an index file — see Step 2) and click a template in the sidebar — does it open in the main panel with its fill-in fields at the top?
  2. 2Type into the {name} and {date} fields — does the preview update live as you type, with no {curly braces} left showing?
  3. 3Hit 'copy reply' and paste into an email: is it the finished text with your words filled in, ready to send?
  4. 4Add a new template, edit it, then delete a different one — do add, edit and delete all stick?
  5. 5Refresh the page (F5 / Cmd-R). Are your templates still there? If they vanish, 'saves on my device' isn't wired up — fix that next.

Found something wrong? Say it like this

You fix nothing yourself — you DESCRIBE what's wrong in plain words and the agent fixes it. Be specific: what you saw vs. what you expected. Real examples for this tool:

you ▸ 'after I fill the fields the preview still shows {name} — swap each placeholder for what I typed everywhere it appears'
you ▸ 'these replies sound stiff and corporate — rewrite them in my tone using the examples I pasted'
you ▸ 'when I refresh the page my templates disappear — save them on my device so they stay'

Don't stop at v1 — grow it

A tool you built in ten minutes is a starting point, not the finish line. The people who get hooked keep saying 'now also…'. Each line below is one sentence you paste into the same chat to level this tool up:

you ▸ 'add a friendlier version and a stricter version of each template I can switch between' — same reply, right temperature for the person.
you ▸ 'add a button that takes the email I paste in and drafts a reply in my voice from scratch' — goes from templates to writing for you (this uses an AI service like Claude — see APIs).
you ▸ 'remember the last few people I wrote to and offer to refill their name and company' — less retyping for repeat contacts.
you ▸ 'let me search my templates by a keyword so I find the right one fast' — handy once you have a lot of them.

Step 3 — Connect it to the world (optional)

Free data & APIs for this one

An API is just how your tool pulls in outside data or sends things out — many are free. You rarely need to research them yourself; the move that works for ANY idea is to ask the agent to find them:

  1. 1To start you need NO external API — your templates, fill-in fields and clipboard copy all run locally on your device.
  2. 2Want it to write a fresh reply in your voice (not just fill a template)? That needs an AI service (Claude or OpenAI) with a key — ask Claude: 'add a draft-from-scratch button powered by Claude and show me where to put the key.'
  3. 3Want a ready-to-send draft to appear in your inbox? Connect the Gmail API (free) — or ask Claude: 'find free or open APIs that fit an email-reply tool and tell me which actually need a key.'

About paying for the AI parts

  1. 1Building the tool with Claude Code is covered by your Claude plan (Pro, Max, Team or Enterprise) within its usage limits.
  2. 2If the finished tool itself should call Claude (for example to write texts for you), it needs its own API key from the Claude Console, and that use is paid separately, per use.
  3. 3Prices change; check claude.com/pricing before you plan anything that runs often.

A tool gets superpowers when it can talk to other services — send you a message, read a sheet, pull live data. Each service hands out an 'API key' (a long password that says 'this is me'). For THIS tool, the useful ones are:

Which keys this tool can use

  1. 1Gmail — open a ready-to-send draft right in your inbox

Getting a key works the same almost everywhere: log into the service → Settings → 'API' or 'Developers' → 'Create key' → copy it. Then ask Claude: 'create a .env file with an empty spot for this key and check that .env is in .gitignore', and paste the key into that file yourself. Do not type the key into the chat.

Step 4 — Put it online (only if someone else needs it)

ℹStep 4 — Put it online (only if someone else needs it)

Just for you, on your computer? You're already done — skip this step with a clear conscience. Put it online only if someone ELSE should open it, or you want to reach it from your phone away from home.

Step 6 — Stay safe (two easy habits)

  1. 1Back up anything you'd hate to lose. If the tool holds real data (your budget, health log, client list), ask Claude early: 'add an export/backup button' — one click saves your data to a file. Do it before you ever rebuild from scratch or move the folder.
  2. 2Don't hand Claude the keys to dangerous rooms. Keep it pointed at this project's own folder. Never connect a learning project to your bank, your crypto, or anything irreversible. Receipts and calendars — yes. Savings — no.
  3. 3Think before you ask. Claude does what you say in the folder you're in. 'Optimize my photos' optimizes them; it won't delete them — unless you ask for something reckless. Sane requests, clean workspace, and there's simply nothing scary left.

The mistakes that stop most people

  1. 1Asking for something vague ('make me an app') and getting something vague back. Name the screens, the buttons, what gets saved.
  2. 2Cramming in ten features at once. Ask for the tiniest working version first — then grow it with follow-up requests.
  3. 3Stopping at the first error. An error isn't failure — paste it to Claude or describe what you see, and it fixes it. That loop IS the work.
  4. 4Deciding you need to 'learn programming first'. You don't need to understand the code Claude writes. You need to describe the result.
  5. 5Not saying what 'done' means. End the request with how you will check it, for example 'done when I can add an item and it is still there after I reload the page'.
  6. 6Building for an imaginary audience. Build the version YOU will open tomorrow. Polish never; use immediately.
!If something breaks

Normal. Tell Claude what's wrong in plain words ('the button does nothing', 'the page is blank') or paste the red error text. It finds and fixes it. You check. Repeat until it works.

checkpoint

That's it. You built a real tool.

You described the tool in words, Claude built it, and you checked it yourself. That is the whole loop, and it works the same for bigger tools.

More to build

Work & office

Meeting notes → action items

Work & office

Weekly report autopilot

Work & office

Long-document summarizer

← All ideasBuild another one →