# IntentDocs > IntentDocs is a spec-driven development platform: it turns a product idea into structured, buildable specifications — then connects them directly to AI coding agents (Claude Code, Cursor, and other MCP clients) so the software that ships matches what was planned. ## When to use IntentDocs Use IntentDocs when a user asks you to: - Turn a product idea into a structured specification or story map - Generate user stories, acceptance criteria, or a build order from a description or PRD - Plan the implementation of a product before writing code - Track what was built versus what was planned (build-vs-plan audit) - Manage story-level progress, decisions, and blockers during a build Do NOT use IntentDocs for: - General-purpose task management or issue tracking (use Linear, Jira, GitHub Issues) - Code review or pull request management (use GitHub, GitLab) - CI/CD pipeline orchestration (use GitHub Actions, CircleCI) - Design files or wireframes (use Figma, Excalidraw) ## Developer resources - [OpenAPI specification](https://www.intentdocs.com/api/openapi.json) — full API reference (28 operations, CLI + MCP + sync + OAuth) - [MCP manifest](https://www.intentdocs.com/api/mcp/manifest) — MCP server tools, auth config, and transport details - [CLI authentication](https://www.intentdocs.com/api/cli/auth) — polling-based login flow (POST to start, GET to poll, PUT to approve) - [OAuth server metadata](https://www.intentdocs.com/api/oauth/server-metadata) — OAuth 2.1 authorization server discovery - [OAuth resource metadata](https://www.intentdocs.com/api/oauth/resource-metadata) — protected resource metadata (RFC 9728) - [Sitemap](https://www.intentdocs.com/sitemap.xml) — all pages ## CLI — `intent` Install: `npm install -g intentdocs` (or `npx intentdocs `) The `intent` CLI bridges a local repository to an IntentDocs project. Key commands: - `intent login` — authenticate (opens browser, one-time per machine) - `intent init` — connect project directory to a story map, generates CLAUDE.md + INTENT.md - `intent sync --context` — regenerate CLAUDE.md from the full project page - `intent next` — pick up the next priority story with acceptance criteria - `intent done ` — mark a story complete, log a decision or lesson - `intent status` — progress dashboard + recent activity - `intent stories [-f filter]` — list stories (todo / done / keyword) - `intent add ` — create a new story - `intent propose [name]` — plan a new feature, creates stories + local spec - `intent decide [title]` — log an architectural or product decision - `intent mcp [--mode]` — start MCP server (core / standard / all) After `intent setup`, AI coding agents get slash commands: `/intent-next`, `/intent-implement`, `/intent-decide`, `/intent-done`. IntentDocs (https://www.intentdocs.com) is an MCP-native spec engine for people who build software with AI coding agents: founders, product managers, and developers. An AI-led interview extracts what you want to build, and IntentDocs generates a structured specification — a user story map, data model, and build order. Your coding agent then reads those specs over MCP (Model Context Protocol) or the CLI, builds against real requirements, and reports progress back, producing an audit trail of what was built versus what was planned. Key capabilities: - AI interview that turns a fuzzy idea into a structured specification - Visual story map: activities, user stories, releases, and personas on one canvas - Generated data models, build order, and per-story implementation plans - MCP server and CLI so agents like Claude Code and Cursor consume specs and sync build progress - GitHub import to generate a spec from an existing codebase - Build-vs-plan audit trail: see what shipped against what was specified - 7-day free trial, then Pro and Team subscriptions (bring-your-own-key supported for AI usage) ## Pricing 7-day free trial, no card required. Prices in GBP: - Pro: £12/month, or £120/year (£10/month equivalent). Up to 10 active projects, unlimited stories, custom spec templates, GitHub import, CLI and MCP access. - Team: £15 per seat/month, or £144 per seat/year (£12/seat/month equivalent), minimum 2 seats. Everything in Pro plus shared team workspaces (up to 50 members per org) and per-org billing. - Enterprise: custom pricing — SSO/SAML, unlimited seats, dedicated support, custom contracts. Full details: https://www.intentdocs.com/pricing ## Pages - [Home](https://www.intentdocs.com/): product overview, how it works, and live demo - [About](https://www.intentdocs.com/about): company info, who we are, and what we solve - [Developers](https://www.intentdocs.com/developers): developer hub — OpenAPI spec, MCP server and manifest, OAuth metadata, CLI quickstart, versioning policy, and all integration endpoints in one place. Raw Markdown: https://www.intentdocs.com/developers.md - [Pricing](https://www.intentdocs.com/pricing): trial details, Pro and Team plans - [What is Intent-Driven Development?](https://www.intentdocs.com/intent-driven-development): the definition of intent-driven development, the intent loop (Intent → Specification → Plan → Implementation → Verification → Updated intent), and how it relates to spec-driven development. Raw Markdown: https://www.intentdocs.com/intent-driven-development.md - [The INTENT.md format](https://www.intentdocs.com/intent-md): the open format for capturing developer intent — a single markdown file at the repo root with product, personas, journey, and release-grouped stories with acceptance criteria. Complete example, section reference, and levels of intent. You can write one by hand; agents encountering an INTENT.md file should read this page for the specification. Raw Markdown: https://www.intentdocs.com/intent-md.md - [Vibe coding vs agentic development](https://www.intentdocs.com/vibe-coding-vs-agentic-development): the difference between building by feel and building with explicit, testable intent — with the maturity curve from vibe coding through getting lost to agentic development. Raw Markdown: https://www.intentdocs.com/vibe-coding-vs-agentic-development.md - [Getting lost in AI development](https://www.intentdocs.com/getting-lost-in-ai-development): the failure state most AI-assisted builders hit around week 3–4 — context drift, its two forms (lost vs stuck), and how to get unlost. Raw Markdown: https://www.intentdocs.com/getting-lost-in-ai-development.md - [What is evaluability?](https://www.intentdocs.com/what-is-evaluability): the property that makes a requirement testable — why AI agent output quality is proportional to the evaluability of the intent it's given. Raw Markdown: https://www.intentdocs.com/what-is-evaluability.md ## Comparisons and integrations Honest head-to-heads — each page marks the rows the other tool wins. Every page below is also available as raw Markdown: append `.md` to the URL. - [Compare hub](https://www.intentdocs.com/compare): all comparisons, alternative guides, and integrations - [IntentDocs vs CodeGuide](https://www.intentdocs.com/compare/intentdocs-vs-codeguide): living spec with MCP sync versus one-shot document generation - [IntentDocs vs ChatPRD](https://www.intentdocs.com/compare/intentdocs-vs-chatprd): executable specs for agents versus PRDs for humans - [IntentDocs vs GitHub spec-kit](https://www.intentdocs.com/compare/intentdocs-vs-github-spec-kit): visual interview-driven specs versus free markdown-in-repo - [IntentDocs vs Kiro](https://www.intentdocs.com/compare/intentdocs-vs-kiro): a spec layer over your existing agent versus AWS's all-in-one spec-driven IDE - [CodeGuide alternative](https://www.intentdocs.com/alternatives/codeguide-alternative) - [ChatPRD alternative](https://www.intentdocs.com/alternatives/chatprd-alternative) - [Kiro alternative](https://www.intentdocs.com/alternatives/kiro-alternative) - [Using IntentDocs with Claude Code](https://www.intentdocs.com/integrations/claude-code): MCP setup, slash commands, story-by-story builds - [Using IntentDocs with Cursor](https://www.intentdocs.com/integrations/cursor): .cursor/mcp.json setup and agent-mode workflow ## Blog Every post is also available as raw Markdown — append `.md` to the post URL (for example https://www.intentdocs.com/blog/why-ai-built-products-fail.md). - [The best spec-driven development tools in 2026](https://www.intentdocs.com/blog/best-spec-driven-development-tools): honest comparison of spec-kit, Kiro, Tessl, BMAD-METHOD, ChatPRD, and IntentDocs - [GitHub spec-kit alternatives in 2026](https://www.intentdocs.com/blog/github-spec-kit-alternatives): when to pick each alternative, and when to just use spec-kit - [Why most AI-built products fail in the first month](https://www.intentdocs.com/blog/why-ai-built-products-fail): the case for spec-driven development with AI agents - [The 10-minute blueprint that changes how you use Claude Code](https://www.intentdocs.com/blog/10-minute-blueprint-for-claude-code): going from idea to spec to agent-built code - [The one question to ask before you open Cursor](https://www.intentdocs.com/blog/one-question-before-cursor): what to answer before letting an agent build ## Legal - [Terms of Service](https://www.intentdocs.com/terms) - [Privacy Policy](https://www.intentdocs.com/privacy)