Overview
1. Log in and choose a workspace
Open app.guidetake.com, log in, and create or join the workspace that should own the guide. Workspace membership controls who can manage the guide and its published content.
2. Create or capture a guide
Create a blank guide in the web app when you want to build the steps manually. To document a real browser workflow, connect the GuideTake Chrome extension, choose Record a guide on Home (or Create → Record a website in the Content library), and select a recordable tab from the current Chrome window or enter a web address. The extension captures pages, clicks, screenshots, and browser context as structured steps. In an existing unlocked Guide, choose Record more and select the exact step after which the new recording should be inserted.
To reuse existing material, open an unlocked Guide and choose Import steps. You can select steps from another Guide in the same workspace, upload as many as 30 JPEG, PNG, or WebP screenshots at 5 MB each, or upload a PDF or DOCX file up to 10 MB. Screenshot drafts can be reordered, renamed, given instructions, and assigned a hotspot before import. PDF and DOCX text is converted into editable drafts; when OPENROUTER_API_KEY is configured, GuideTake also structures those drafts with AI, and otherwise keeps the deterministic extracted result. Choose the target step or the end of the Guide before completing the import.
Workspace owners and admins can open Workspace settings → Customization before creating guides to set defaults on five tabs: Appearance (default view, or scroll-only and walkthrough-only guides, page layout, brand color, font, hotspots and right-to-left text), Scroll View (navigation, initial zoom and image scaling), Walkthrough View (step text, step counter, zoom, mobile layout, talking-head placement, autoplay and the call to action), Sharing (PDF download and embeds) and Feedback (reactions and comments). Changes apply only after Save, and leaving the page with unsaved changes asks for confirmation. Options your plan does not include show a crown and an upgrade link instead of an editable control. New guides inherit these defaults whether they start blank, from text, from a video or from an extension recording. In a guide's Customize panel, Apply defaults copies the current workspace defaults again while keeping that guide's sharing settings; each guide can still be adjusted later.
3. Review every step
Open the guide editor to reorder steps and refine titles, descriptions, hints, branches, and calls to action. Generate an editable guide title or summary from the ordered steps when you want a starting draft; nothing is persisted until you review and save it. The FAQ editor can customize its public heading, append grounded questions generated from the same guide content, reorder them, or remove the complete list.
Choose Edit text under a step's instructions to add headings, bold, italic, strikethrough, lists, and links. AI paraphrase creates a draft that you can review before choosing Save text. Press Ctrl/Cmd+Enter to save or Escape to cancel. Failed saves keep the draft; switching steps or switching to voiceover text keeps it within the current editor. Save before leaving or reloading the page. Existing plain instructions keep literal punctuation, while saved formatting is carried into public playback, translation, copying and HTML/Markdown exports.
Outside text fields, press Ctrl+K on Windows/Linux or Cmd+K on macOS to open Quick actions. Inside the rich-text editor this shortcut edits a link. Search Quick actions for a setting or choose Jump to to find a step by its title or instructions. Outside text fields, Shift+J jumps to steps, Shift+I opens imports, Shift+B opens branding, Shift+X opens exports, and Shift+E opens embed settings.
Choose Sort steps, select the steps you want to move, and drag a handle to move them together. Shift-click selects a contiguous range; Move up and Move down offer keyboard and touch alternatives. Choose Save order to persist the draft order or Cancel to leave it unchanged. If another editor or an extension recording changes the guide meanwhile, reload the latest order and apply your changes again.
Use the checkboxes beside steps to select a batch for deletion. Deletion requires confirmation and cannot be undone. Branches pointing to removed steps return to automatic navigation. Forms attached at those steps are detached from this guide; the reusable forms and their existing submissions remain available.
The selection toolbar also offers Copy and Hide media. Copy preserves the saved steps as they were when copied, including their media, translated text, and ready voiceovers. Open a guide in the same workspace and choose Paste steps from an insertion menu, or press Ctrl/Cmd+V outside a text field to insert after the current step. You must still have access to the source guide and edit access to the destination. A changed destination requires reloading before retrying; pasted steps keep internal branches but references to unselected steps return to automatic navigation. Guide-level settings, comments, media history, and form attachments are not part of the step clipboard.
Ctrl/Cmd+C copies the selected steps when you are not editing or selecting text. Ordinary text copy/paste continues to work inside text fields. Hide media temporarily collapses the editor preview to text; it does not remove media from copied steps or published guides. Clear the selection or press Escape to show media again.
A step can upload a JPEG, PNG, GIF, WebP, or MP4 file directly to private object storage, attach a direct HTTPS image or video, or embed a YouTube, Vimeo, or Loom link. Uploaded objects are verified before the step is updated, and viewers receive time-limited playback links instead of storage keys. For an uploaded PNG or JPEG screenshot, choose Redact sensitive information, select the detector categories, and review the newly stored redacted image before publishing. This feature requires the deployment's Google Cloud Sensitive Data Protection configuration; the original private media object remains available to the storage lifecycle instead of being overwritten.
4. Add language and voice options
When your audience needs them, translate guide content between English and Simplified Chinese, use AI paraphrase to tighten instructions, and add a voiceover to each step. Voiceover can come from a verified audio upload, a browser microphone recording of up to 30 seconds, or generated speech. Each step keeps its narration script separate from displayed text; when that script is blank, generated speech reads the displayed instructions. Review text and audio before publishing.
On paid plans, AI actions such as rewriting, translation, generated voiceover and document or video import use the workspace's monthly AI credits. Each action shows the credits it needs before it runs, and a request that fails is not charged.
By default, voiceover uses Fish Audio's official voices. The voice list recommends voices for the narration language and starts from the voice this guide already uses; the Browse all official voices entry pinned at the bottom of the list opens the full library, where you can search, filter, listen to samples and star favorites. Any voice can narrate any language, and the list notes when a voice mainly speaks another language. On the Business plan, the list also offers Clone a voice…: record 10 to 30 seconds of clear speech in a quiet room, or upload 10 to 60 seconds of audio, then name the voice and confirm that it is yours or that you have the speaker's permission. Everyone who edits guides in the workspace can then narrate with it. A workspace keeps up to 10 cloned voices; cloning uses no AI credits, and narration in a cloned voice costs the same as other generated voiceover. GuideTake keeps the sample so the voice can be rebuilt, and deletes it together with the voice; narration already generated keeps its audio. On the Business plan, a workspace owner or admin can connect the workspace's own provider keys under Workspace settings → Integrations: an OpenRouter key with a model, picked from a short list of mainstream models or entered as an OpenRouter model ID, for text AI features, Help Center Ask AI and document imports; a Fish Audio key for voiceovers in official voices; and a Groq key for Video-to-Guide transcription. GuideTake checks each key with its provider and encrypts it at rest. Each connected key has a Use this key first switch. Off (the default), AI features spend this period's AI credits first and switch to that key only when credits run short; on, they use the key first. Video-to-Guide uses both its OpenRouter and Groq keys, so it uses them first only when both switches are on. Work that runs on your keys uses no AI credits and is billed by the provider, and an AI action shows which keys it will use before it runs. Cloned voices and screenshot PII redaction always use AI credits. If a provider refuses a key or its balance runs out, the action stops without charging credits. The deployment must set a unique INTEGRATION_CREDENTIALS_ENCRYPTION_KEY with at least 32 characters before a key can be saved.
5. Publish and share
Open Share in the guide editor to use the Access, Modes, Link, Embed, and Export options. A public published guide with link sharing enabled can be opened by anyone with its link. For a private guide, invite a person by email with view or edit access; they must sign in to GuideTake using that email address. Workspace access can separately allow all members to view or edit the guide.
In Share → Modes, control Share via link and Share via embed independently. Turning links off still permits allowed embeds; turning embeds off leaves enabled links available. Both switches save automatically. A failed save leaves the saved settings unchanged; if another editor changed the guide, reload the current settings before retrying.
Enable Domain restriction to limit embedding to listed domains and their subdomains. Enter domains or HTTP(S) URLs separated by commas; the field saves when you leave it. URLs are normalized to their hostname, not a path, protocol, or port. A restricted empty list blocks all embeds. Missing embedding referrers and disallowed ancestor websites also block loading. Lists support up to 100 entries and must fit the browser policy header; shorten an oversized list before enabling that restricted policy. Existing oversized lists remain editable even while embedding is off.
Sharing modes control delivery, not confidentiality. They do not grant access to private guides, revoke downloaded content, or prevent a non-browser client from forging browser request headers. Use private-guide permissions for sensitive content. An embedded viewer left open for more than 12 hours may need to reload to renew its delivery context; changing the embedding policy invalidates existing contexts for subsequent requests.
Use Share → Embed to choose Scroll, Slides, or Walkthrough (when the guide has visual media). These options belong to the individual embed snippet; they do not change the guide's saved default view. Slides includes an introduction and compact previous/next controls. Walkthrough can skip the introduction and show controls on hover or persistently; portrait media also offers an orientation choice. The preferences menu controls flexible or fixed 640 px sizing, rounded corners, and a border. Fixed embeds shrink to fit a narrower host.
For Scroll, choose Parent height to scroll inside the host's frame, or Guide height to expand the frame to its contents. Guide height includes a small script served by GuideTake; the destination site must allow both the iframe and this script. Messages are checked against the iframe's exact origin and window. Hosts that remove scripts can use Parent height instead. Copy the updated snippet whenever you want to change an existing embed.
The inline preview and Preview embed use your signed-in collaborator access and display the saved guide, including saved drafts. They do not create viewer statistics or save form responses, and background music does not autoplay. Public embeds still require publication, enabled embed sharing, the allowed embedding domain, and any private-guide permissions. Private embeds also depend on the destination browser's cross-site cookie policy.
Export creates portable Markdown, HTML, or print/PDF output. Export → Print / PDF opens an authenticated preview of the current saved guide, including drafts, all steps and expanded FAQ answers. Use its print button after it loads. This collaborator-only export remains available with both sharing modes disabled and does not create viewer analytics or require completing guide forms. Background music does not autoplay in the print preview. The separate viewer-download setting still controls the print button on public playback.
Export → HTML offers HTML file only, which uses online media links that may expire, or Offline HTML + media (.zip). The ZIP includes the selected chapters' images, videos, narration and Talking Head clips with the saved text and selected FAQs. Extract it and open index.html, keeping the media folder beside it. Website links still need internet; forms and analytics are not included. The archive is limited to 512 MiB, so select fewer chapters if needed. A failed media download or changed saved content stops the export; check access and try again. Markdown and clipboard copies continue to use online media links.
In Share → Link → Custom guide URL, the workspace owner first sets a Workspace link prefix. A guide editor can then choose a Guide short name, use the guide title as a starting point, and save it. Both parts accept 3–250 lowercase letters, numbers, or hyphens. For example, the prefix acorn and short name invite-your-team produce https://app.guidetake.com/guides/acorn~invite-your-team. This prefix is separate from the workspace's navigation address.
The saved address is used for the copied link and QR code. New embed snippets use the stable guide ID so changing a custom URL does not break an installed embed. A configured custom domain on a published Help Center also provides https://help.example.com/guides/invite-your-team, scoped to that domain's workspace. Publishing and guide access rules still apply; a private guide requires an authorized sign-in, and Help Center password protection does not replace the guide's own permissions.
Changing the workspace prefix changes all of its custom guide URLs. Changing or resetting a guide short name makes its old custom URL stop working; existing stable /guides/<guideId> links and guide content are kept. Addresses remain reserved while a guide is in Trash, and clones start without a custom address. A duplicate name, locked guide or changed revision prevents saving and keeps your input so you can retry.
The Directory also lets you organize Articles in the same folders as Guides, search or favorite them, and review content shared with you. In the Article editor, write Markdown, preview GFM tables and task lists, upload a 1920 × 1080 cover, and insert published interactive Guides. Article sharing supports public or restricted links, workspace-wide access, individual view/edit invitations, and dynamic placeholders supplied through URL query parameters. Export the current Article as HTML or Markdown.
Workspace owners and admins can organize published Guides and Articles in a searchable Help Center. Categories support five nested levels and can be moved by changing their parent; each content item can be assigned or moved independently. After changing published content or access, verify the visitor experience again.
When workspace defaults disable embeds, new guides do not expose embed code until an editor enables sharing via embed for that guide. Link sharing remains independently controlled in Share → Modes.
6. Learn from engagement
Use guide analytics to review unique views, starts, step progress, completions, CTA interactions, reactions, and form submissions. Use those signals to improve steps where readers stop or hesitate. Search engines and other bots are not counted, and neither are teammates who can edit the guide or article when they open its GuideTake link while signed in. Daily totals follow your browser's time zone.