Short answer: Go to Learning, scroll to Learning sources and click Add additional sources. Choose Websites to let replient.ai read your website, or Documents (pdf, etc.) to upload an FAQ, a price list or brand guidelines (TXT, PDF, DOCX or XLSX, up to 5 MB per file). Your past replies on social media are learned automatically when you connect a channel.
Before you start
Website training is included in all plans.
You need the permission to manage learning. The brand owner always has it.
Documents must contain real text. Scans and images are not readable.
What the AI learns from
Source | What it teaches the AI | How it gets in | How it stays up to date |
Your past replies | How you answer: tone, wording, typical answers | Automatically when you connect a channel | Every reply you send or edit is added. Replies from automations and bulk actions are not |
Website | Facts: products, prices, shipping, opening hours | Websites source | Re-read every day |
Documents and FAQs | Facts and ready-made answers | Documents (pdf, etc.) source | Only when you upload a new file |
Learning works across channels. What replient.ai learns from your replies on Facebook, Instagram, TikTok, LinkedIn, YouTube, Google Reviews and your other connected channels is used together for the whole brand, so your voice stays consistent everywhere. Each brand learns separately.
Your own rules for tone and style belong in custom instructions, not in a document. See How do I control the AI's tone, language and style?.
Step 1: Open your learning sources
In the sidebar, click Learning. Scroll down to Learning sources (1) and click Add additional sources (2).
A window opens. At the bottom you find the two source types: Websites (1) and Documents (pdf, etc.) (2).
Step 2: Connect your website
Click Websites.
Enter your domain (1), for example aurora-skincare.example, and click Connect now (2).
The source appears under Learning sources with the label Learning in progress. When the label changes to Active, the first reading is done.
replient.ai finds your pages through your sitemap and refreshes the knowledge every day. Only changed pages are processed again.
Step 3: Upload documents and FAQs
Click Documents (pdf, etc.).
Click the upload area (1) or drag your files into it. You can add several files at once.
Each file shows Reading file…, then Training, then Trained.
Click Connect now (2) to keep the files. If you close the window without clicking it, the uploaded files are removed again.
Supported files
Formats: TXT, PDF, DOCX, XLSX (CSV also works). Maximum 5 MB per file.
Text only: If a PDF is a scan, you see No readable text found. Export it as text first.
A good FAQ file: one question and its answer per row or paragraph. Write the answer the way you want the AI to say it.
Priority: Your past replies on social media take precedence over documents if they contradict each other.
How long until the AI uses it?
Documents: as soon as the file shows Trained. Training a file can take a minute.
Website: it starts right after you click Connect now and runs in the background. The label Learning in progress tells you when it is finished. Large websites take longer.
Past comments: after you connect a channel, replient.ai reads your earlier comments and replies. This usually takes a few minutes, depending on volume.
New knowledge is used for every new suggestion. Comments that already have suggestions keep them. To get a fresh set, click the circular arrow below the suggestions (free of charge).
My website can't be connected
Check these points in this order:
Sitemap: replient.ai needs a sitemap to find your pages. It looks for a Sitemap: line in your robots.txt, then for /sitemap_index.xml and /sitemap.xml. Without one you see: Please add a sitemap to your robots.txt so we can check all your pages. Most website builders create a sitemap for you; add its address to your robots.txt.
Firewall or bot protection: The message Our request was blocked by your server means your site refused the request. This is common with Cloudflare. Follow Allow replient.ai to access your cloudflare protected website. Your robots.txt is only used to find the sitemap. It is not what blocks the connection.
Your own domain: Social media profiles and file-sharing links (Instagram, Facebook, Google Drive, Dropbox and similar) are not accepted. Use your own website address, or upload the content as a document.
Text in the page: replient.ai reads the text that is part of the page itself. Content that only appears after JavaScript runs is not read.
Language folders: replient.ai prefers German pages. If your site has /de pages, or both un-prefixed pages and /en pages, only the /de and un-prefixed pages are read and the /en pages are skipped.
Already trained: The message Already trained means all pages are up to date. Nothing is wrong. The same website cannot be added twice (Source already exists).
Still blocked? Contact us in the chat and send us the address.
Remove or update a source
Go to Learning → Learning sources.
Click the three dots on the source card and choose Delete. The website or file is removed from the AI's knowledge.
To update a document, delete it and upload the new version. A website updates itself every day.
⚠️ Important: Only delete sources in the Learning sources list here. If you delete a connected social media account in the list above, replient.ai also deletes its comments and everything it learned from your past replies on that channel.