Already have an account? Sign in →Free 14-day trial · No credit card required

Ship the right thing,
not just something.

Fettink turns plain-language requirements into a full spec stack — structured requirements, tech spec, system design, and task specs — with team approval gates between every stage. From PM idea to sprint board in one workflow.

Get started free See how it works

See it in action

From requirements to shipped PR in one structured flow.

Choose your workflow

Three workflows for three sizes of work. Vibe for small changes. Spec-Driven for features that need team alignment. Design-Driven for multi-week, multi-engineer features.

Vibe

Bug fixes, quick changes. AI generates an implementation brief.

Spec-DrivenSelected ✓

Requirements → Tech Spec → Design → Tasks, with approval gates.

Design-Driven

Design doc decomposed into parallel blocks with a roadmap view.

Three workflows. One size doesn't fit all.

Match the process to the scope of the work. Over-engineering a bug fix wastes time. Under-speccing a major feature ships the wrong thing.

Vibe

Write fast, ship now

Bug fixes, config changes, quick refactors. Describe what needs to happen — AI generates a focused implementation brief. No gates, no ceremony.

Best for: Scope is obvious, risk is low.

Spec-Driven

Team alignment before code

Requirements → Tech Spec → System Design → Task Specs, with team approval gates between each stage. Nothing moves forward without sign-off.

Best for: Standard features that need peer review.

Design-Driven

Multiple engineers, in parallel

A hand-crafted design document decomposed into independent implementation blocks. Each block gets its own mini-spec. A Gantt view shows the full dependency graph.

Best for: Large multi-week features, multiple engineers.

Spec work is a tax. We cut it by 80%.

Most spec work is repetitive: translate PM requirements into engineering language, generate acceptance criteria, write the design doc, break it into tickets. Fettink automates the generation so your team reviews and refines instead of writing from scratch.

Before Fettink

Requirements doc1–2 hrs

EARS acceptance criteria30–45 min

Tech spec + design2–3 hrs

Task breakdown + estimates1 hr

Total4–7 hrs

With Fettink

Requirements doc5 min

EARS acceptance criteria1 click

Tech spec + design10 min

Task breakdown + estimates1 click

Total~30 min

~2.5 hrs

saved per spec session

$250+

value per session at $100/hr eng cost

1 session

to recover a month of Pro cost

Everything from idea to sprint board

Every other spec tool is single-player and stops at the document. Fettink covers the full chain — from raw PM requirements to ticket-level tracking — with team review baked in.

4-stage spec pipeline

Requirements → Tech Spec → System Design → Task Specs. AI generates each stage from the prior one. Review and edit at every step before moving forward.

Team approval gates

Configure required reviewer count per stage. Reviewers approve or request changes. Nothing advances until all required sign-offs are collected.

EARS acceptance criteria

AI generates structured acceptance criteria in Easy Approach to Requirements Syntax (WHEN/THEN, WHILE/SHALL, IF/THEN). Every requirement is testable and auto-posted to linked GitHub PRs.

Sprint board with story points

AI decomposes task specs into Fibonacci-pointed sprint tickets. Drag tickets through To Do → In Progress → In Review → Done. GitHub PRs auto-transition columns on open and merge.

PM → engineering handoff

PMs capture requirements and business value in a dedicated PM workflow. AI generates EARS acceptance criteria. One click creates a pre-populated spec session for the engineering team.

GitHub integration

Link PRs to sessions or individual tickets. EARS criteria post to PRs automatically. Tickets move to In Review when a PR opens and Done when it merges.

Global roadmap & Gantt view

A cross-project view showing every active session, workstream, and timeline in one place. Visualize dependencies and estimated delivery dates across your entire engineering queue.

Analytics dashboard

Track AI token usage, session volume, and generation latency over time. See usage by provider and per-seat token consumption at a glance.

Input & output guardrails

Block PII, profanity, and custom restricted terms before they reach the AI. Output guardrails validate that generated content meets your standards before it's shown.

Team library

Share prompt templates and spec sessions across your org. Fork any completed session to start a related feature with inherited requirements and EARS criteria already in place.

Spec anchors

Tag named sections of your spec documents. When the document changes, stale anchors are flagged so downstream dependencies (code, tests, docs) stay in sync.

Intent-to-prompt

Describe your intent in plain language and let AI clarify the goal, scope, and constraints before generating. Reduces back-and-forth by catching ambiguity upfront.

Enterprise

Brand voice enforcement

Configure tone, style, preferred vocabulary, and terms to avoid. All AI-generated specs and documents apply your brand voice automatically.