SPECLAN v0.9 — First Public Release
SPECLAN ships its first public release. A VS Code extension that turns your specifications into AI-ready prompts — powered by Claude Code.
SPECLAN v0.9 — First Public Release
Today we're shipping SPECLAN v0.9 — the first public release of our specification language and VS Code extension.
Your Specification Is Your Prompt
AI coding agents have become remarkably capable. The bottleneck is no longer the agent — it's the input. The better the specification, the better the code. SPECLAN is the missing middle layer between what your business needs and what your AI agent builds.
Write a one-sentence feature description. Get a full requirement tree back. Hand it to Claude Code. Ship.
What's in v0.9
This release ships with a complete specification workflow — from first draft to tested, released code:
- Hierarchical Domain Model — Goal, Feature, Requirement, Scenario, Acceptance Criterion, Test. Nothing exists in isolation. Nothing falls through the cracks.
- WYSIWYG Markdown Editor — Business people write in a polished editor. Developers see clean Markdown. Same file, same tool, no translation layer.
- 7-Status Lifecycle — Draft, review, approved, in-development, under-test, released, deprecated. Always know where every requirement stands.
- AI-Powered Specification Generation — Describe a feature in a sentence. The AI Generation Wizard builds a full specification tree in minutes — requirements, scenarios, acceptance criteria, tests.
- HLRD Import Assistant — Feed in a stakeholder brief, a product requirements doc, or meeting notes exported as Markdown. Get structured goals, features, and requirements back.
- Codebase Inference Assistant — Already have code? Point SPECLAN at your project and reverse-engineer a full specification tree from what's built.
- Implementation Assistant — Once a spec is approved, the Implementation Assistant crafts an implementation plan ready to hand off to Claude Code.
- MCP Integration (42+ Tools) — Your specifications become accessible to any AI assistant that supports the Model Context Protocol.
- Git-Native Collaboration — Branch, merge, and review specs the way you already collaborate on code. No database. No server. Just Markdown files in your project folder.
- Change Requests — Released specs are locked. Changes go through their own lifecycle — tracked, reviewed, implemented. No silent spec drift.
Built for Claude Code
SPECLAN is built exclusively for Claude Code. That's a deliberate choice, not a limitation.
Claude Code is the most capable AI coding agent available today. Rather than building lowest-common-denominator integrations across every AI tool, we went deep on one. SPECLAN uses the Claude Agent SDK under the hood. The AI assistants — generation, clarification, implementation planning, codebase inference — all run through Claude's agentic capabilities. The result is an integration that feels native, not bolted on.
Setup takes one command. If you have a Claude subscription — Max, Pro, or Team — open a terminal in VS Code, run claude login, and every AI feature lights up. No API key. No separate billing. No token budgets to manage. You already pay for Claude. SPECLAN puts that investment to work inside your specification workflow.
No Lock-In. No Paperwork. No Excuses.
Every specification lives as a plain Markdown file with YAML frontmatter in your project's speclan/ directory — right next to src/. The folder structure is the data model. Stop using SPECLAN tomorrow and your specs remain perfectly readable. No export step. No migration script. No proprietary format.
speclan/
├── goals/ G-###-slug.md
├── features/ F-####-slug/F-####-slug.md
│ └── requirements/ R-####-slug/R-####-slug.md
└── templates/
Any person — or any AI agent with file system access — can read and understand your specifications just by looking at the file system.
Get Started
- Install SPECLAN from the VS Code Marketplace
- Open a terminal in VS Code and run
claude login - Create your first specification — or import one
New project? Import a requirements doc. Existing codebase? Infer specs from code. Either way, you have structured, AI-ready specifications in minutes.
Your ideas deserve structure. Your specification is your prompt. Let's build.