Skip to content

FAQ ​

Short answers to the questions site owners, administrators and developers ask most, with links to the page that explains each one in full.

10 minutesEveryoneSingle user · SaaS

Recording ​

Does Replora record a video of the screen?

No. The tracker records the page's structure (a DOM snapshot) and every change to it, plus mouse movement, clicks, scrolling and typing (masked), with the open-source rrweb recorder. The replay rebuilds the page from that. It is much smaller than a video, text stays sharp at any size, and anything masked was never captured at all. See The tracking script.

Does the tracking script slow down my site?

It loads with async, so it never blocks rendering or the page's load event. It is about 43 KB gzipped, sends a compressed batch about every five seconds, and every part of it is wrapped so an error in the tracker can't break your page. See Size and performance.

Which third parties receive my visitors' data?

None by default: the script, the rules and the recordings all go to your own Replora server. Optional features you switch on talk to other services: AI summaries and Ask AI send session moments to Anthropic (Claude), the IPinfo location provider receives visitor addresses, HubSpot receives replay links for identified contacts.

Are passwords and card numbers recorded?

Never. Everything typed is masked by default (***), and password fields, payment card fields (autocomplete="cc-…") and fields whose name says they hold a secret stay masked even when you allow typed values. See Privacy & masking attributes.

How do I hide personal data shown on a page?

Add class="sr-mask" (or data-sr-mask) to mask an element's text, or class="sr-block" (or data-sr-block) to leave it out entirely. Without touching the site, list CSS selectors under Mask the text of these elements and Leave these elements out in the website's settings. See Privacy.

Does it work with React, Vue, Angular and other single-page apps?

Yes. The tracker watches pushState, replaceState, popstate and hashchange and counts each address change as a page view. The npm packages @replora/browser and @replora/react add router hooks for Next.js, React Router, Remix, Vue, Nuxt and SvelteKit. See Single-page apps & frames.

Can one visit across two of my domains be one session?

Yes: put the same website's snippet on both domains and push __sr.push(['crossDomain', ['shop.example', 'pay.example']]) on both. Links between them then carry the session. See Cross-domain sessions.

Are iframes recorded?

Frames of the same domain are recorded with the page. Frames of another domain appear as empty boxes unless you switch on Record frames from other domains and put the same snippet into the framed pages. See Frames and iframes.

Are native mobile apps recorded?

The tracker records websites and web apps, in desktop and mobile browsers alike. Native iOS and Android apps are not recorded. (The REST API is what a mobile client of the dashboard uses.)

Troubleshooting ​

I installed the snippet but no sessions appear. What should I check?
  1. Open a page of your site with ?sr_diagnostics=1: the panel says whether recording runs and why not, and what the server answered.
  2. Allowed domains must include the domain the page runs on (a 403 in the panel means it doesn't).
  3. Recording enabled must be on in the website's settings (a 404 means it is off or the key is wrong).
  4. The background worker must run: batches wait in the queue until php artisan queue:work processes them. See Background worker.
  5. Check the website's rules: sampling, Only record sessions that last at least, start pages, excluded IP addresses (your office?) and targeting.
  6. You may have opted your own browser out (?sr_optout=1); ?sr_optout=0 undoes it.
Replays show the page without its styles or images. Why?

The replay loads the page's stylesheets and images from your site. If they are behind a login, on localhost or a private address, or blocked for other origins, the replay can't load them. Keep Keep copies of stylesheets on (the server keeps copies of linked CSS), and use Load images and styles from another address in replays to point replays at a public copy of your assets. See Troubleshooting.

My site has a Content Security Policy. What do I need to allow?

Your Replora server in script-src and connect-src. Surveys and the A/B loader also need style-src 'unsafe-inline'. See Content Security Policy.

Many visitors share one office or mobile-carrier address and some aren't recorded.

Tracker requests are limited per IP address: 600 a minute by default, about 50 active visitors behind one address. Raise RECORDINGS_INGEST_PER_MINUTE (or the same setting in Settings › Installation). See Environment.

How do I exclude my own and my team's visits?

Open the Exclude your own visits link from the Install page (?sr_optout=1) once in every browser your team uses, or list your office's IP addresses under Never record these IP addresses.

How do I record only after consent (GDPR, ePrivacy)?

Load the snippet only after the visitor agreed, or set window._srOptOut = true before it loads for visitors who declined. __sr.push(['optOut']) stops a running recording and keeps the browser opted out. See Consent and opting out.

Does Replora set cookies?

Not by default: the tracker uses sessionStorage and localStorage of your site. It sets one first-party cookie (_sr_v) only when you choose One visitor across subdomains of. The full list is under Browser storage.

Are IP addresses stored?

Only shortened: the last part of IPv4 addresses (the last 80 bits of IPv6) is zeroed before anything is stored, and Anonymize visitor IP addresses more strongly keeps only the first two parts. Excluded addresses are compared before storing and never kept.

Is Do Not Track respected?

Not automatically. If your policy requires it, set window._srOptOut = true when navigator.doNotTrack === '1' or navigator.globalPrivacyControl is true, before the script loads.

Storage and data ​

Where are recordings stored?

In storage/app/recordings on the server by default, or in S3-compatible storage (Amazon S3, Cloudflare R2, MinIO) with RECORDINGS_DRIVER=s3. Session lists, statistics and reports are in the PostgreSQL database. See Recording storage.

How long are recordings kept?

Per website: 7, 30, 90, 180 or 365 days (Keep recordings for), capped by the plan in the SaaS edition. Expired recordings are deleted every night at 03:00. See Plan limits.

Which database does Replora need? Can I use MySQL?

PostgreSQL 13 or newer. MySQL and MariaDB are not supported. Redis is optional (cache, sessions, queue).

Editions, licensing and updates ​

What is the difference between the Single user and SaaS editions?

The Single user edition runs Replora for you and your team. The SaaS edition adds public sign-up, plans with limits, Stripe or Paddle billing, a public website with its own pages, an affiliate program and a support desk, so you can sell Replora to customers. REPLORA_MODE switches between them; SaaS needs the saas/ folder. See Editions.

Do I need a purchase code?

Yes for regular releases: the installer checks it with the license server and downloads the modules it brings, signed. The SaaS edition accepts only codes of the SaaS plan. See Your purchase code.

How do updates work? Can I go back?

The System page shows a new release when one is out; Update now downloads it, checks its signature, keeps your current files as a backup and updates the database. php artisan replora:update --rollback puts the previous files back. Your .env, recordings and own modules stay. See Updating & rolling back.

Can I run Replora on shared hosting?

Yes, with PHP 8.3 and PostgreSQL. Use the web installer, set up the cron line for the scheduler and a way to run the queue worker; where the host forbids background programs, QUEUE_CONNECTION=sync processes recordings during the request (for small sites). See Shared hosting.

Developers ​

How do scripts and other servers access the data?

Through the REST API under /api/v1 with a token from Developers › API tokens (full access or limited to recordings:read / recordings:write). See REST API.

How do I know a webhook really came from my Replora server?

Each alert webhook carries X-Replora-Signature: sha256=<HMAC-SHA256 of the body> made with the alert's secret. Verify it over the raw body. See Verifying the signature.

Can I attach the replay to a support ticket from my own app?

Yes: __sr.push(['getSessionLink', function (link) { … }]) hands your code the dashboard link to the visitor's running session. See getSessionLink.

Can I add my own features?

Yes, as a module: a folder in modules/ with module.json, a service provider, routes, migrations and dashboard pages. See Writing a module.

AI, languages, scale ​

What do the AI features need, and what do they send?

An Anthropic API key (ANTHROPIC_API_KEY or Settings › Installation). A summary sends the session's moments (pages, clicks, errors, masked as recorded, at most 600 of them) to Claude. Summaries, searches and Ask AI answers are made when someone asks; a site summary, once an account manager switches it on, is rewritten every few hours. Plans can limit summaries and Ask AI questions per month. See AI features.

Which languages does the dashboard speak?

English, Spanish, French, German, Portuguese, Italian, Russian, Chinese, Japanese, Korean, Turkish, Indonesian and Arabic (right to left). Administrators can change any text. See Languages & translations.

Do I need WebSockets for the live view?

No. By default the dashboard polls every 5 seconds. For instant updates run Laravel Reverb and set REPLORA_REALTIME=websocket.

How far does it scale?

The ingest endpoint only checks and queues batches, so the web servers stay light; the work is done by queue workers, which you can add as traffic grows. With several servers use Redis or the database for cache and sessions, S3 storage and a separate ingest queue. php artisan replora:doctor checks the setup. See Scaling to high traffic.

Replora: see what your visitors see.