Specification-Driven Development

Your system is born across dozens of AI chats.Specky turns them into one coherent specification.

Describe what you want to build. Specky asks questions, detects contradictions, and records every decision — who, when, why. The result: a versioned system plan for your dev team, or the input for Claude Code to build it step by step.

7 days free · no card · your specs never train AI models.

The problem

Chat is where your project is decided.
It's also where your project gets lost.

History scattered across chats

The architecture in one chat, the data model in another, a scope change in the fifth. Two weeks later, nobody knows where the current version lives.

No trace of decisions

"Why did we pick PostgreSQL over Mongo?" The answer is somewhere in the scroll. A chat doesn't remember who decided, when, or what was rejected.

AI gets incomplete context

You paste fragments from memory into your agent. Code gets built from what you remember — not from what was actually decided.

Chat is great for thinking. Terrible as the source of truth about your project.

The layer chats are missing

You already have a chat to think in.
Specky is where the project becomes real.

  ChatGPT / Claude / Gemini Specky
Decision history scattered across chats one project, one source of truth
Conversation context you paste fragments the model sees your whole project specification
Versioning none specification snapshots with history for every requirement
Decisions lost in the scroll who, when, why — saved on the requirement
Consistency the model doesn't see the whole contradiction detection across requirements
Teamwork copy-paste links shared project, roles, joint resolution
Agent connection none — copy-paste MCP server: the agent reads the spec, proposes changes, reports progress
Implementation state "did we do that?" status of every requirement: open → code ready → done / to fix

Specky is the layer chats are missing — it completes your stack, it doesn't replace it.

01 · Provenance

Every decision has an author and a reason.

You resolve a question — Specky records the decision on the requirement: who, when, what the alternative was. The requirement's history shows how it reached its current shape. Come back a month later and you know why, not just what.

Requirement 3.2 · history3 versions
v3export format: PDF → PDF + CSV
Anna · resolves conflict with 5.1 · 2 days ago
v2added criterion "max 100k rows"
Marek · 5 days ago
v1created from spec.pdf · §4
import · 2 weeks ago
02 · Consistency

Contradictions surface before they reach the code.

Requirement 3.2 says "export to PDF". Requirement 5.1 says "CSV only". A chat won't catch it — it doesn't hold the project as a whole. Specky points to both sources and asks you to resolve it.

AI proposes, the human decides. No change without your approval.

Contradiction detected
Req 3.2 — export to PDFsource: spec.pdf · §4
— conflicts with —
Req 5.1CSV onlysource: email · Mar 14
Keep PDF
Keep CSV
Allow both
Nothing changes until you resolve it.
03 · MCP connector

Your agent, connected to the spec.

One command and Claude Code reads requirements straight from Specky — with acceptance criteria and full context. The agent can also propose changes: they go to your review queue, never straight into the spec.

When the agent finishes an implementation, the requirement gets status code ready. You test and mark it done — or to fix with a note the agent picks up as a task. Specky always knows what already exists in the code.

The agent reports it's finished. You're the one who says done.

terminalMCP
$claude mcp add --transport http specky https://specky.app/mcp
opencode readydone
↳ or to fix — note sent back to the agent
04 · Meeting intake

A whole meeting turns into reviewed spec changes.

Drop in a call transcript or your raw meeting notes. Specky reads the whole thing and works out where it belongs in the project — new requirements, agenda points to decide, acceptance criteria — each proposal pointing back to what was actually said.

Nothing enters the spec on its own. You approve each proposal, edit it, or dismiss it.

standup-2026-07-14.txt · 48 min6 proposals
requirementOffline mode for the mobile app"…it has to work on the metro with no signal" · 12:04
agendaWhich regions get offline support first?open question for the client · 24:31
criterionSynced within 5s of regaining signalon Req 4.2 · 25:10
Approve
Edit
Dismiss
Every proposal waits in your review queue.
Who it's for

Two ways to build with Specky.

Solo builder with AI

You + Claude Code / Cursor

Specky is the memory and discipline your agent lacks: one coherent spec instead of forty chats, markdown export, MCP in your terminal.

  • Idea description → requirements with acceptance criteria
  • Clarifying questions before any code exists
  • Anthropic's latest models included — talk through your system inside the project, no API keys or separate subscriptions
  • claude mcp add — the agent reads the spec, no copy-paste
  • Input for Claude Code: spec export + starter prompt
  • Change a feature → get a diff: what to add, what to fix
Small IT team

Agency · startup · software house

Requirements, client decisions, and choices in one place — with roles and attribution. No more "who decided this, and where".

  • Joint resolution of questions and contradictions
  • Client-decision agenda linked to requirements
  • Drop in meeting notes or transcripts — Specky proposes requirement updates for your review
  • Roles: owner, client, developer, viewer
  • Delivery-plan export (markdown / PDF) for developers
  • A decision history readable by non-devs too
How it works

One loop, from idea to shipped — and back.

01

Describe

Start a conversation about your project — like ChatGPT or Claude, but focused on your spec. Or upload docs (PDF, DOCX, MD). Requirements emerge as proposals: you accept or reject each one.

02

Clarify

Specky asks questions and flags contradictions between requirements. Talk each one through — with your full spec as context, not a pasted fragment. The moment you decide, it's saved as a decision on the requirement.

03

Freeze a version

A coherent set of requirements with acceptance criteria, captured as a snapshot. Whoever asks — team, client, or agent — you always know which version is current.

04

Hand off & come back

Export the plan, or feed Claude Code (spec + starter prompt). The agent ships requirement by requirement and reports code ready; you mark done or to fix. New feature → diff → next version.

The specification doesn't end at v1. Specky carries it through the whole life of the project — and knows what of it already exists in the code.

Specky completes your stack — it doesn't replace it. Chat to think Specky as the source of truth Your agent to code
Trust

Your project stays yours.

Data hosted in the EU (Hetzner). Your files stay on EU servers under GDPR.

Your specs never train AI models. Paid APIs only, with training excluded.

Every AI proposal is justified — with a cited source you can check.

FAQ

Questions, answered.

How is this different from writing a spec in Claude Projects / ChatGPT?
Same brains, different layer: Specky runs on Anthropic's latest models too. The difference is everything around the conversation — versioning, decisions with attribution, contradiction detection, team collaboration, and MCP for your agent. Chat thinks — Specky remembers and guards consistency.
How is this different from .md files in a repo (spec-kit)?
Files have no API. Through MCP the agent asks Specky for the current requirements, and its proposals go through your review queue. Plus a decision history readable by non-devs too.
Does Specky write code?
No. Specky produces a coherent specification; your team or an agent (e.g. Claude Code) writes the code from it.
Do I have to use Claude Code?
No. Export the delivery plan to markdown / PDF — it works for humans just as well. MCP and agent input are an option for those working with AI.
I already have documentation — what about it?
Upload it (PDF, DOCX, MD). Specky extracts requirements with a pointer to the spot in the source, and asks questions where the description is ambiguous.
What about my data?
EU (Hetzner), paid APIs with no training on your data, and a full change history.

Start your next project with a specification that won't get lost in chats.

7 days free. No card.