Book a call
HomeServicesPortfolioAboutContactBlogCareers
Book a call
Free tools

AI technical spec generator

Describe your product idea in plain words. In about a minute you get a first-draft technical specification: user roles, MVP and later features, data, integrations, non-functional requirements, the questions to answer before development, and a ballpark in hours. Free, no sign-up.

Try an example: 0 / 2000
What are you building?
Starting point
Payments
Takes about a minute. Up to 3 specs a day. Don't include passwords or personal data.

What you get: an example

Generated from this idea: A marketplace for freelance tutors. Parents search tutors by subject, price and availability, book a lesson and pay online. The platform keeps a commission and pays tutors out weekly. Tutors manage their calendar and see their earnings. Both sides get email and SMS reminders, and parents can leave a review after a lesson.

Freelance Tutor Marketplace

A web marketplace where parents find tutors by subject, lesson rate and availability, book lessons and pay online. Tutors manage their calendars and earnings, while the platform deducts a commission and schedules weekly payouts.

Audience: Parents booking tutoring for their children, freelance tutors and marketplace administrators.

User roles

  • Parent: Search and compare tutors; Book and pay for lessons; View and cancel own bookings; Review completed lessons
  • Tutor: Manage profile, subjects and lesson rates; Set availability and calendar exceptions; View and manage own lessons; Complete payout onboarding; View earnings and payout status
  • Administrator: Approve or suspend tutor profiles; Manage subjects and commission settings; Resolve booking and payment issues; Moderate reviews; Monitor payouts

Features

First release (MVP)
  • Accounts and permissions. Provide email-based accounts with parent, tutor and administrator permissions and ownership checks.
  • Tutor profiles. Publish tutor biographies, subjects, lesson rates and profile photos after administrator approval.
  • Tutor discovery. Filter approved tutors by subject, lesson rate and bookable availability.
  • Availability management. Let tutors define recurring availability and exceptions using explicit time zones.
  • Booking and checkout. Temporarily reserve a lesson slot during checkout and confirm it only after verified payment without allowing overlapping bookings.
  • Cancellations and refunds. Apply agreed cancellation rules, release cancelled slots and track refund outcomes.
  • Commission and weekly payouts. Record the commission and tutor share for each payment and issue weekly payouts for eligible completed lessons.
  • Tutor earnings dashboard. Show tutors completed lesson earnings, deductions, pending balances and payout history.
  • Email and SMS notifications. Send booking confirmations, cancellation notices and scheduled lesson reminders to parents and tutors.
  • Verified lesson reviews. Allow one parent review per completed booking and provide administrator moderation.
  • Marketplace operations. Provide an administrator panel for tutor approvals, booking support, refunds and failed payout investigation.
Later
  • External calendar synchronization. Synchronize tutor bookings and external busy periods with a connected calendar to reduce scheduling conflicts.

Data

  • User: Account identity, role, contact details, time zone and notification preferences.
  • TutorProfile: Public biography, subjects, lesson rate, photo and approval status.
  • Subject: Administrator-managed subject taxonomy linked to tutor profiles.
  • AvailabilityRule: Recurring bookable windows and dated exceptions in the tutor's time zone.
  • Booking: Parent, tutor, subject, lesson interval, temporary reservation expiry and lifecycle status.
  • Payment: Provider references, payment and refund states, and the commission policy captured at checkout.
  • LedgerEntry: Immutable records of tutor earnings, platform commission, refunds and payout allocations.
  • Payout: Tutor, weekly batch, allocated earnings, provider reference and settlement or failure status.
  • Review: Booking-linked rating, review text and moderation status.
  • Notification: Recipient, channel, scheduled send time, delivery status and deduplication key.
  • CalendarConnection: Tutor's external calendar authorization, encrypted credentials and synchronization cursor.

Integrations

  • Marketplace payments and tutor payouts (Stripe Connect or Adyen for Platforms): Choose a provider supporting the launch countries, tutor identity verification, commissions and weekly payouts; verify webhook signatures and reconcile payments, refunds, disputes and payout failures idempotently.
  • Transactional email (Amazon SES, Postmark or SendGrid): Send confirmations and reminders asynchronously; process bounce and delivery webhooks and avoid duplicate sends.
  • SMS reminders (Twilio or Vonage): Validate phone numbers, respect consent and messaging restrictions, handle delivery callbacks and provider rate limits.
  • External calendars, later phase (Google Calendar or Microsoft Outlook): Use OAuth for booking export and busy-time import; handle token revocation, webhook expiry, incremental synchronization and API limits.

Background jobs

  • Expire unpaid booking reservations and release held slots.
  • Send email and SMS reminders in each recipient's time zone.
  • Mark elapsed lessons as awaiting completion or dispute resolution according to the agreed policy.
  • Calculate weekly payout eligibility, submit payouts and reconcile settlement results.
  • Retry transient notification failures and reconcile missed payment webhooks.
  • Synchronize connected calendars and renew subscriptions in the later phase.

Non-functional requirements

  • Booking integrity: Enforce non-overlapping tutor bookings atomically, including simultaneous checkout attempts and daylight-saving transitions.
  • Financial reliability: Use immutable financial records, idempotent payment processing and reconciliation so retries cannot duplicate charges, refunds or payouts.
  • Security: Enforce record ownership, protect administrator actions, encrypt integration credentials and keep card data with the payment provider.
  • Privacy: Minimize children's personal data, restrict access to contact details and define retention and deletion rules that preserve required financial records.
  • Operations: Use managed hosting, automated deployments, database backups and alerts for failed payments, overdue payouts and reminder delivery failures.

Suggested stack a suggestion, not a decision

  • A backend framework with transactional APIs and a built-in administrative interface.
  • A relational database such as PostgreSQL for bookings, permissions and financial records.
  • A durable job queue and scheduler, optionally backed by Redis.
  • An object storage service compatible with S3 for profile photos.
  • A web application framework with server-rendered tutor pages and authenticated parent and tutor dashboards.
  • Managed cloud infrastructure with CI/CD, centralized logs and error monitoring.

Assumptions

  • The initial release supports one-to-one lessons with a fixed duration and no recurring booking packages.
  • The marketplace launches in one country with one settlement currency.
  • Parents hold accounts; children do not sign in or communicate directly through the platform.
  • Lesson delivery happens outside the platform; built-in video and chat are excluded.
  • Tutors use provider-hosted identity verification and payout onboarding.
  • The commission policy is centrally configured and captured on each booking.
  • Only completed, undisputed lessons are eligible for weekly payouts.
  • External calendar synchronization is deferred until after the initial release.

Open questions to answer before development

  1. Which launch country and currency must the payment provider support, and who is legally responsible for collecting payments and handling disputes?
  2. What are the cancellation, rescheduling, tutor absence and parent no-show rules, including refund eligibility?
  3. How is lesson completion confirmed, and how long may a parent dispute a lesson before its earnings become payout-eligible?
  4. How is commission calculated, and who bears payment processing fees, refunds and chargebacks?
  5. What happens if a refund or chargeback arrives after the tutor has already been paid?
  6. Are lessons online, in person or both, and must discovery include location or travel distance?
  7. Are lesson durations fixed, and are buffers, booking notice periods or recurring lessons required?
  8. What tutor checks are required before publication, particularly where tutors work with children?
  9. Can parents manage multiple children, and what child information must tutors see?
  10. What booking volume, tutor count and peak concurrent checkout demand should the first release support?

Why start with a spec

Every development team asks for a spec before it can give a real estimate. Without one you get either weeks of back-and-forth or a number pulled out of thin air. A spec does not need to be long: it needs to say who uses the product, what goes into the first release, what data it keeps, which services it talks to, and which decisions are still open.

This generator gives you that first draft. It does not replace a conversation with an engineer, but it makes that conversation much shorter, and the open questions show you what you have not thought about yet.

How it works

  1. You describe the idea. A few sentences in plain words are enough: who it is for, what they do, what the product must handle.
  2. AI drafts the spec. A language model, instructed by our backend engineers, structures it into roles, features, data, integrations and requirements, and lists its assumptions instead of guessing silently.
  3. The calculator adds hours. The AI only picks which parts of the work your idea needs. The hours come from our backend estimate calculator, where you can adjust the scope.
  4. You take it anywhere. Copy it, download it as Markdown, or send it to us for a free review by an engineer.

FAQ

What is a technical specification for software?

A technical specification (spec, or technical requirements document) describes what a product must do before anyone writes code: who uses it, which features go into the first release, what data it stores, which services it connects to and what it must handle in terms of load and security. Development teams estimate and plan from it.

How does the AI spec generator work?

You describe your idea in plain words and answer three quick questions. An AI model, instructed by our backend engineers, turns that into a structured first draft: roles, MVP and later features, data, integrations, non-functional requirements, assumptions and open questions. It takes about a minute.

Is it free? Do I need to sign up?

It is free and needs no account or email. To keep it free for everyone, each visitor can generate up to 3 specs a day.

Can I use the spec with any development team?

Yes. Copy it or download it as a Markdown file and send it to any team or freelancer. The suggested stack is deliberately language-agnostic, and nothing in the spec ties you to us.

How accurate is the hours estimate?

It is a ballpark for the first release (MVP); the features marked "later" are shown as a separate total. The AI only picks which parts of the work your idea needs; the hours come from our backend estimate calculator, where every item has a typical range for an experienced team. Open it in the calculator to adjust the scope. A real estimate needs an engineer to go through the open questions with you.

What should I do with the open questions?

Answer them before you ask for quotes. They point at decisions that change the scope the most, such as refunds, permissions or who owns the data. A team that gets the answers up front can give a much tighter estimate.

What happens to the idea I describe?

Your description is sent to an AI provider (OpenAI) to generate the spec. API providers do not use this data to train their models. We keep the description and the spec for 30 days so the result link keeps working, then delete them. Don't include passwords or personal data in the description.

Rather talk it through?

ISTRALLEN builds backends, payment and API integrations, and production AI features, and helps teams fix and deploy the systems they already have.

Book a call Estimate a backend project