# IntentDocs vs CodeGuide (2026) *Comparison · [Read on intentdocs.com](https://www.intentdocs.com/compare/intentdocs-vs-codeguide)* > Choose CodeGuide for one-shot breadth — a batch of documents with wireframes, 40+ models, and a Chrome extension into web builders. Choose IntentDocs when the spec needs to survive the build: your agent reads it over MCP and syncs progress back, giving you an audit trail of built versus planned. Disclosure: we build IntentDocs, so this comparison has a side. The table below marks the rows CodeGuide wins. ## What is CodeGuide? CodeGuide turns a plain-English idea — or an imported GitHub repo — into a batch of AI-ready documents: project brief, PRD, tech-stack doc, wireframes, project plan, and a task list. You paste or sync those into Cursor, Windsurf, Claude Code, or web builders like Lovable, helped by a Chrome extension and six open-source starter kits. Its pitch is spec-driven context: front-load the context and the agent's first attempt works. ## What is IntentDocs? IntentDocs is a spec-driven development platform: an AI interview turns your idea into a structured spec — a visual story map, data model, and build order — and your coding agent (Claude Code, Cursor, or any MCP client) builds from it over MCP, with an audit trail of what was built versus what was planned. ## Feature by feature | Feature | IntentDocs | CodeGuide | Our call | | --- | --- | --- | --- | | Getting the spec out of your head | AI interview asks what you're building, who it's for, what done looks like | You write a prompt or fill a form; it generates the document batch | IntentDocs | | Output breadth | Story map, data model, build order with per-story plans | PRD, tech-stack doc, wireframes, project plan, task list | CodeGuide | | Wireframes and design prototypes | Not offered | Included on paid tiers | CodeGuide | | Model choice | Anthropic models, bring-your-own-key supported | 40+ LLMs on paid tiers | CodeGuide | | Working inside web builders (Lovable, Bolt) | No extension — IntentDocs targets agent workflows over MCP/CLI | Chrome extension syncs context straight into web builders | CodeGuide | | Connection to your coding agent | MCP server and CLI — the agent reads the spec and writes progress back | Copy/paste, one-command attach, or extension sync — one-way handoff | IntentDocs | | After the build starts | Living spec: audit trail of built-versus-planned as the agent works | Documents are static once generated; drift is invisible | IntentDocs | | Spec from an existing codebase | GitHub import | GitHub import | Tie | | Pricing model | Flat subscription from £12/month | Credit-metered from $24/month | IntentDocs | ## Pricing **CodeGuide** (August 2026): - **Free:** $0 — 120 one-time credits, 1 project, 5 LLM models - **Pro:** $24/mo ($288/yr) — 5,000 monthly credits, up to 15 projects/month, 40+ models, wireframes, GitHub import - **Founder:** $29/mo ($348/yr) — 6,000 monthly credits, bring-your-own-keys, priority support, early access CodeGuide is credit-metered: generation draws down a monthly credit pool, with paid top-ups beyond it. **IntentDocs:** 7-day free trial, no card required. Pro is £12/month (or £120/year, equivalent to £10/month). Team is £15 per seat/month (or £144/seat/year), minimum 2 seats. Enterprise is custom. Full details: https://www.intentdocs.com/pricing ## Choose CodeGuide when… - You want wireframes and design prototypes generated alongside the docs. - You build mainly in web builders like Lovable or Bolt and want a Chrome extension injecting context where you work. - You want to choose between 40+ models rather than one vendor's. - You want starter kits tuned to the specs it writes, and the reassurance of a large existing user base. ## Choose IntentDocs when… - You want the spec to stay alive during the build — your agent reads it over MCP and syncs progress back, so you can see built-versus-planned instead of hoping. - You think in user journeys: a visual story map you can walk a co-founder through beats a folder of generated documents. - You'd rather have an interview extract the product from your head than prompt-engineer a generator. - You want flat, predictable pricing instead of watching a credit meter. ## The honest verdict CodeGuide is the stronger one-shot generator: more document types, more models, wireframes, and a distribution story (extension + starter kits) that meets web-builder users where they are. IntentDocs is built on a different bet — that the hard part isn't generating the first spec, it's keeping spec and code connected once the agent starts building. If you want documents to hand off, choose CodeGuide. If you want a spec your agent builds against and reports back to, choose IntentDocs. ## Frequently asked questions **Is CodeGuide better than IntentDocs?** For generating a broad batch of documents — PRD, tech stack, wireframes — with your pick of 40+ models, yes, CodeGuide is ahead today. For keeping a spec connected to the build — agent reads it over MCP, syncs progress back, and you get an audit trail of built-versus-planned — IntentDocs is the stronger choice. **Can both generate a spec from an existing codebase?** Yes. Both CodeGuide and IntentDocs can import a GitHub repository and generate specification documents from the existing code. **Do both work with Claude Code and Cursor?** Both target them, differently. CodeGuide hands documents off — paste, one-command attach, or extension sync. IntentDocs connects over MCP: the agent pulls stories and context directly and writes build progress back to the spec. **Which is cheaper?** IntentDocs Pro is £12/month flat (£120/year). CodeGuide Pro is $24/month with a 5,000-credit meter. If you generate heavily, flat pricing is easier to predict; if you generate rarely, CodeGuide's free tier (120 one-time credits) may be all you need. --- *Facts about CodeGuide verified August 2026 against their published pricing and documentation (https://www.codeguide.dev).* *Alternative guide: https://www.intentdocs.com/alternatives/codeguide-alternative · Full tool roundup: https://www.intentdocs.com/blog/best-spec-driven-development-tools*