A recurring check is not a guarantee of publication. Your saved schedule starts a publishing opportunity; source availability, approvals, access, media validation and destination limits still apply. For example, an unchanged RSS feed can produce no post, while an unchanged AI prompt can deliberately produce a new variation.
Open all six modules in the Auto Post dashboard. Each source below has its own full setup guide.
What each source does
| Source | What happens on a scheduled occurrence | How to control unnecessary work |
|---|---|---|
| AI Prompt | Generates a fresh variation from your saved brief. The prompt does not need to change. | Choose a frequency that matches how often you want new content. Pause campaigns you no longer need. |
| YouTube Video | Reads available metadata and captions and prepares a fresh variation. Captions are not always available; metadata alone does not justify invented video details or quotes. | Use this for recurring reuse of one video. A video URL is not a subscription to all future videos on a channel. |
| Blog URL | Reads the public article and prepares a fresh variation, even if the article has not changed. | Use RSS for new-article monitoring. Use a blog URL when you want to reuse a particular article. |
| RSS or Atom feed | Finds up to one new item per recurring publishing window. Unchanged items skip AI text and image generation. Optional publisher notifications can wake an unused window sooner, with 15-minute backup checks. | Review RSS notifications and check frequency. A 15-minute schedule plans 96 checks per active workflow each day. |
| Document | Prepares the next unused record. Ready-made content avoids AI text rewriting; the rewrite mode uses AI. When every record is used, the occurrence skips instead of retrying as a source failure. | Replace the file to continue, or pause a finished workflow to stop future checks. It does not automatically import new records from a file on your computer. |
| Ecommerce | Refreshes the connected catalog or public product page and selects an eligible product for the next promotion. The same product can be promoted again in a later rotation. | Choose relevant products, review availability rules and set a sensible frequency. Connected catalog requests return up to 100 products; use the source search/filter if a selected product is outside those results. |
What Postly reuses
Preparation for the same saved recurring occurrence is reused across its destinations and retries. Postly checks source-item history before generating AI text or images for an unchanged RSS item or a used document record. This helps avoid preparing the same occurrence separately for every channel.
Overlapping requests for the same exact public article or video metadata URL can share an in-flight download. After that request finishes, those article and product-page responses are not retained for the next occurrence. Private connected store catalogs remain scoped to their workspace and connection and are refreshed through their authorized provider clients.
RSS has a separate, bounded short-lived cache with conditional requests when the publisher supports them. This reduces downloads for unchanged feeds; it does not eliminate network requests or guarantee a single fetch across all servers. Public source downloads have a 2 MB limit. Uploaded documents have a 2 MB file limit, up to 500 records, and up to 8,000 characters per record; review the parsed preview before saving.
Review your actual schedule and costs
Open the workflow's publishing settings to review its recurring schedule, start time, time zone, end date and any separate schedules for individual channels. Existing workflows keep their saved settings until you change them. RSS notifications do not turn every Auto Post source into an instant event trigger.
AI text, generated images and normal destination publishing can use different allowances. Ready-made or Publish feed content avoids AI text rewriting, but selected AI media can still use credits. An unchanged prompt, article, video or product can intentionally generate another paid variation on its next scheduled run. Use previews, review the first result and choose a slower schedule when you do not need frequent variations.
Paused workflows do not publish. A genuine source outage follows the normal bounded retry policy. An exhausted document or unchanged RSS item is an intentional skip. Check workflow history and channel results before deciding to retry a failed or uncertain publication; retrying blindly can create duplicates.