Octopus.do
Octopus.do is a browser-based visual sitemap builder for planning websites and their content: start from scratch, crawl an existing site, or generate a whole page hierarchy from an AI prompt. No sign-up is required to begin.
What is Octopus.do?
Octopus.do is a browser-based tool for planning the architecture of a website before anyone writes a line of code. Its own definition is direct: a web-based tool that creates visual website sitemaps from a short text description, using AI to generate page hierarchies, suggest content structure, draft text for each content block and produce SEO meta tags, replacing the manual work usually done at the earliest stage of web planning. Work happens on an artboard. Pages sit in a tree and each page can carry content blocks. A project opens three ways: from scratch, by pasting a URL into the Crawler so an existing site is imported with its meta tags, or by writing a prompt and letting the AI produce the hierarchy. The publisher calls its approach the Content brick method and presents it as the middle ground between a mind map and a wireframe. Each page holds stackable lo-fi wireframes, combined with Shift and click, that show what the page will contain without pretending to be a design. Around that core sit the artboard elements: web and mobile page frames, arrows for internal links and user flows, Symbols managed globally so one edit propagates everywhere, status tags, sticky notes, external links and attachments. Two dedicated modes finish each page, SEO mode for Title, Description, H1 and notes, and Content mode with a multi-style text editor. The AI assistant generates pages, drafts block copy from the project context, produces SEO tags, changes colors, checks spelling and translates. The result leaves the tool in many shapes: PDF, PNG, Excel, CSV, XML, TXT, a submittable sitemap.xml, a view-only embed, plugins for Figma, WordPress and Webflow, and a Publish function that turns the plan into a simple responsive website. Unlimited editors are included on every plan, with comments, access rights and assign or resolve. Machines can drive it too, through a public REST API and an OAuth 2.1 MCP server. The publisher is SIA Scada, based in Riga, Latvia, with a product team described as two people. The domain was first captured by the Wayback Machine in April 2019, and the site claims 150,000 users.
What it does
- Generate a complete sitemap from a description of a few words
- Crawl an existing website, 300 pages on the free plan and 3,000 on paid plans, to extract its structure and meta tags
- Stack lo-fi wireframes on each page to show what its content will be
- Draft and rework block copy with AI, adjusting tone, length, spelling and language
- Generate the SEO tags (title, description, H1, notes) for an entire project
- Export to PDF, PNG, Excel, CSV, XML and TXT, or push to Figma, WordPress and Webflow
- Drive the whole project from an AI agent over MCP, or from code over the public REST API
When to use Octopus.do / When not to
A quick filter to help you decide if Octopus.do is the right fit.
When to use Octopus.do
- Web agencies running client discovery and scoping, where one testimonial calls the tool the backbone of that process and describes drawing sitemaps live with the client
- UX/UI designers and information architects who need a lo-fi structure agreed on before opening Figma
- Freelancers and small studios planning a redesign, who can crawl up to 3,000 existing pages and restructure them on one artboard
- SEO specialists and copywriters writing titles, descriptions and H1 tags page by page across an entire project
- Product owners and project managers costing a build, using per-page time estimates multiplied by an hourly rate
When not to use Octopus.do
- Anyone looking for a full website builder: the Publish function outputs a simple, document-like site, not a production website
- Designers who need high-fidelity mockups, since the wireframe library is deliberately lo-fi and hands the work over to Figma
- Buyers who insist on testing a paid plan before paying, as there is no trial of Pro or Team, only a refund window
- Teams that need a mobile app or a browser extension, the product being browser-only
- Organizations expecting a localized interface or round-the-clock support, from an English-only product whose product team is described as two people
How to use Octopus.do
A typical end-to-end flow, from setup to results.
- Open the site in a browser: no account is needed to start
- Choose a starting point, from scratch, from a template, by pasting a URL into the Crawler, or by writing a prompt for AI Generate
- Shape the tree on the artboard, adding and moving pages and grouping them into sections
- Drop lo-fi wireframes onto each page, stacking several types together with Shift and click
- Fill in the copy in Content mode, or let the AI draft each block from the project context
- Switch to SEO mode to write the Title, Description, H1 and per-page notes
- Share the link with view, comment or edit rights, or protect it with a password on the Pro plan
- Export to PDF, PNG, Excel, CSV, XML or sitemap.xml, push to Figma, WordPress or Webflow, or publish the plan as a website
- Optionally connect an AI agent to the MCP server over OAuth, then work by instruction, asking it to show your projects or add a pricing page
- For developers, call the REST API with an X-API-Key header or an OAuth 2.1 flow, trying endpoints in the bundled Swagger UI
Pros & Cons
Pros
- A permanent free plan with no time limit, and a product you can start using without creating an account at all
- Unlimited editors on every plan, the free one included
- Fully public pricing, laid out plan by plan in a comparison table of more than fifty rows
- Three complementary ways into a project: from scratch, by crawling an existing site, or from an AI prompt
- An unusually wide export chain: PDF, PNG, Excel, CSV, XML, TXT, sitemap.xml, Figma, WordPress, Webflow and embed
- A public REST API documented endpoint by endpoint with a Swagger UI, plus an MCP server with per-client instructions that can be revoked from the account
- Unusual transparency on data handling: a dated public subprocessor list with countries, a sourced written commitment that OpenAI API data is not used for training, and documented minimization on the MCP side
Cons
- The GDPR is never mentioned anywhere on the site, legal pages included, although the publisher is established in the European Union
- The registered address diverges between legal pages, Gustava Zemgala gatve 76 on the cookie and privacy pages against Rupniecibas str. 32b on the terms, for the same company and the same registration number
- The legal documents contradict each other twice, a generic Free Trial section in the terms against a flat No in the pricing FAQ, and non-refundable subscription fees against a promised full refund within fourteen days
- No DPA is offered or mentioned, no Data Protection Officer is named, and one mailbox serves legal, privacy and support alike
- No user-side setting to keep your data out of training: the commitment covers the OpenAI API, not a choice you can make
- A tightly constrained free plan, one project, one user, 200 MB, a 300-page crawler and limited wireframes, colors, SEO and AI, with no way to test the paid plans first
- Gaps around the product: no mobile app or browser extension, no about page and no named team, no list of the languages the AI translates into, and Do Not Track explicitly unsupported
Pricing & Plans
Octopus.do offers a permanent free plan with no time limit, covering one user, one project, unlimited editors, 200 MB of storage and a 300-page crawler. The lowest paid entry point is the Pro plan at USD 14.00 per month on annual billing, that is USD 168.00 per year, against USD 16.00 per month on monthly billing. The Team plan is USD 20.00 per seat per month on annual billing, against USD 24.00 monthly, and an Enterprise tier is quoted on request. Annual billing is advertised as a 15% saving. No trial of the paid plans is offered; the pricing page states that a full refund may be requested within fourteen days of the first payment.
- one user
- one project
- unlimited editors
- 200 MB
- a 300-page crawler
- limited wireframes
- colors
- SEO and AI
- one user
- eight projects
- 10 GB
- a 3
- 000-page crawler
- branding
- Symbols
- Figma sync
- unlimited projects
- 25 GB
- everything in Pro plus a team workspace
- project folders
- admin roles
- assign and resolve
- restricted access
- anonymous editing
- aimed at larger companies with complex needs and controls
- with advanced security tooling
- no dedicated plan
- but discounts for educational institutions granted on request by email
Data, GDPR & hosting
A consolidated view of how Octopus.do handles your data.
GDPR overview
The site never uses the words GDPR or General Data Protection Regulation. A full-text search across every legal page (privacy policy, terms, cookie policy, subprocessor list) returns no occurrence, and no compliance claim is made anywhere. That is silence, not refusal, because the framework in place is European in construction. SIA Scada is established in Latvia, inside the EU. The privacy policy uses the vocabulary of Data Controller, Data Processor and Data Subject, states rights of access, rectification, erasure and portability, and declares that data is transferred to Latvia and processed there. A dated list of nineteen subprocessors with their countries was published on 9 September 2024. Against that: no DPA is offered or mentioned, no Data Protection Officer is named, and a single address handles every request. No Article 27 representative applies, the publisher being itself EU-based. The stated minimum age is 13.
Who owns the data?
SIA Scada, the Latvian publisher, declares itself the Data Controller of the personal data collected through the service. Ownership of the rest is split. The terms state that the Content you create remains yours and that you warrant holding the rights to it, while intellectual property in the platform itself belongs to SIA Scada and its licensors. Payment card details are never stored by the publisher and pass straight to the payment processors, PayPal/Braintree and Stripe. Users may request access and a copy of their data, rectification, erasure and portability, and the publisher may verify identity before acting on a request. The privacy policy in force is dated 20 September 2023.
Reuse rights
The content you produce is yours to reuse without asking permission: you warrant that you hold the rights to it, and the tool is built to let it leave, through PDF, PNG, Excel, CSV, XML, TXT and sitemap.xml exports and through the Figma, WordPress and Webflow plugins. On its own side, SIA Scada uses data to provide and maintain the service, notify changes, run interactive features, handle support, analyze and improve the platform, monitor usage, detect incidents and send offers and news you can unsubscribe from. The AI sitemap generator collects data including the project ID, to run the feature, inform platform improvements, and meet security and legal requirements. The terms commit that data is never sold or shared for advertising or marketing. Inputs and outputs passed to the OpenAI API are stated not to be used for model training, though the publisher declines responsibility should OpenAI change its policy. Over the MCP server, data leaves only when you ask for it inside your assistant, never in the background and never on a schedule: your account display name, your index of projects, workspaces and folders, and the full content of open projects. Other members' names, emails, photos, roles, permissions and assignments are stripped before sending, as are your own email, plan, billing details and project previews. Cookies cover essentials, preferences and analytics, and Do Not Track is not supported.
Data retention & training
Hosting summary
Three hosting subprocessors are named on the dedicated page: Digital Ocean and Google Cloud in the United States, and Nano IT in Latvia. The privacy policy states that data, personal data included, is transferred to Latvia and processed there, which puts the declared jurisdiction inside the European Union while part of the infrastructure sits in the United States. DNS resolution matches that split, the domain answering on 35.184.225.56, an address located in Council Bluffs, Iowa, on Google's AS396982. What the site does not offer matters as much. There is no selectable hosting region, no data residency promise, and no hosting certification cited anywhere, neither SOC 2 nor ISO 27001. Payment cards are never stored by the publisher: PCI-DSS handling is delegated to PayPal/Braintree and Stripe.
Things to keep in mind
Risks and trade-offs to weigh before adopting Octopus.do.
- The AI can invent content: the terms warn that the generator may make up facts, including the visual sitemap itself and page and section titles, and that nothing it says should be relied on without double-checking
- The legal file contradicts itself in three places, two different registered addresses, a Free Trial section against a pricing FAQ that denies any trial, and non-refundable fees against a promised full refund within fourteen days, which weakens all three as commitments
- Privacy governance is thin: the GDPR is never named, no DPA and no Data Protection Officer exist, and a single mailbox handles legal, privacy and support requests
- You cannot opt your own data out of training: the no-training commitment covers the OpenAI API, and the publisher explicitly declines responsibility if OpenAI changes its privacy policy
- Leaning on generated structures can quietly erode the habit of thinking an information architecture through; a sitemap that looks finished is not a sitemap that has been reasoned about, and a client shown one in a workshop may take it as settled
- The product team is described as two people, a real dependency for a tool placed at the center of an agency process
- Do Not Track is not supported, and the .do domain sits under the Dominican registry although the company is Latvian, which misleads anyone inferring jurisdiction from the domain name
Setup & Integrations
Technical difficulty
Very low for ordinary use. Octopus.do opens in a browser with nothing to install and no account required to start, and testimonials describe instant value with nothing to explain. Two optional layers raise the bar. Connecting an AI agent to the MCP server takes a few minutes: add the URL as a connector and complete an OAuth flow, with no key to manage and instructions given per client. The Figma and WordPress plugins install on the third-party side. Only the REST API needs developer skills, with an X-API-Key header or an OAuth 2.1 flow, and a Swagger UI for testing.
Deployment
Integrations
Supported languages
Behind Octopus.do
Social
Resources
All the official URLs gathered for verification and reference.
Frequently asked questions
Do I need an account to try Octopus.do?
Is there a free trial of the paid plans?
Does the free plan expire?
What exactly does the AI do?
Can I start from a website that already exists?
Which export formats are available?
Can I connect an AI assistant, or write code against it?
Is my data used to train AI models?
Who publishes the tool and where is the data held?
How do I reach support, and is there a minimum age?
Should you pick Octopus.do?
Octopus.do is a mature and unusually well-documented product for its size. A public REST API described endpoint by endpoint, an OAuth 2.1 MCP server with per-client instructions, a changelog running past thirty entries and fully public pricing all point to a team that ships and explains what it ships. The positioning is narrow and openly so: the upstream phase of a web project, structure, content planning, SEO scaffolding and client sign-off, not production. Treated as a website builder it disappoints; treated as the discovery and scoping tool it claims to be, it does the job. The economics suit a freelance or a small agency. The free plan is real rather than decorative, permanent, with no card and no sign-up needed to start, and unlimited editors, while Pro at USD 14.00 per month on annual billing is a modest line item. The weak point is the legal file. It is correct in form, a named Latvian company with a registration number, four dated documents, nineteen subprocessors listed with their countries, yet it never says the word GDPR, gives two different registered addresses on two different pages, carries a Free Trial section in the terms that the pricing FAQ flatly denies, and declares subscription fees non-refundable where that same FAQ promises a refund. None of this is evidence of bad faith; all of it is evidence of documents that were never reconciled. Anyone contracting on behalf of an organization should ask for the current registered address, the refund terms and a data processing agreement in writing first. With those reservations, and with 150,000 claimed users behind a product online since 2019, it remains a solid and honest way to plan a website.
- Choosing a selection results in a full page refresh.
- Opens in a new window.