Loading page
About this tool
Build a clear system prompt for ChatGPT, Claude or Gemini: pick a role such as coding assistant, editor, HR recruiter, product manager or kids tutor and fill in the details.
Pick a role, fill in the fields and copy a ready system prompt with tone, rules and output format. This page brings together 5 separate system prompt builder pages (Coding assistant, Editor, HR recruiter, Product manager, Kids tutor) so you can compare them without hunting for the right one.
Open System Prompt Builder on AltFTool — it loads instantly in your browser.
Choose your role at the top: Coding assistant, Editor, HR recruiter, Product manager and 1 more. On a phone it is a drop-down list.
Fill in the system prompt builder for that role; each of the 5 options keeps its own rules, data and defaults.
Read the result, then switch role to compare. The page link keeps your role (?v=coding-assistant), so you can bookmark or share it.
Coding assistant, Editor, HR recruiter and the rest sit behind one picker, so switching role takes a tap instead of a new search.
Every role keeps its own rules, numbers and defaults rather than one generic template.
The link remembers the role you picked, so a bookmark or a shared link opens straight on it.
5: Coding assistant, Editor, HR recruiter, Product manager, Kids tutor. Pick one at the top of the tool; the page link updates (for example ?v=coding-assistant) so you can bookmark or share that exact view.
Five things: the assistant's role and seniority, the exact stack (language, framework, package manager, test framework), project context, hard rules it must never break, and the output format for code. The most common omissions are naming the package manager — without it assistants mix npm and pnpm commands — and stating whether answers should be diffs, snippets or full files.
Proofreading fixes surface errors only — typos, spelling and punctuation. Copyediting adds grammar, usage and consistency (hyphenation, capitalisation, number style) without touching structure. Line editing goes deeper, reworking sentences for clarity, rhythm and concision while keeping the author's meaning and voice. Each level includes everything below it, which is why this builder writes the chosen level into the prompt explicitly.
In many places yes, but with growing conditions: New York City's Local Law 144 requires an independent bias audit and candidate notice before using an automated employment decision tool, the EU AI Act classifies recruitment AI as high-risk, and Illinois regulates AI video-interview analysis. The safest pattern — and this builder's default — is AI that drafts and structures while a human makes every actual decision. Confirm your jurisdiction's rules with employment counsel.
RICE, created at Intercom, scores each item as (Reach × Impact × Confidence) ÷ Effort. Reach is users affected per period, Impact is rated 0.25 / 0.5 / 1 / 2 / 3, Confidence is 50%, 80% or 100%, and Effort is in person-months. Higher scores rank higher; the value of the exercise is mostly in arguing about the four inputs, which is why this builder makes the assistant show its arithmetic.
Constrain four things in the system prompt: allowed topics (with a rule to redirect everything else), personal data (never ask for names, schools, addresses or photos), sensitive disclosures (one gentle sentence, then refer the child to a parent or teacher), and tone (never mock mistakes). A prompt reduces risk but does not replace supervision — an adult should stay aware of the conversations.