# Scan intervals, pausing and historical backfill (https://snitchfeed.com/docs/listeners/scan-intervals) SnitchFeed collects posts in two ways: * **Real time:** Reddit and Bluesky. New posts are matched against your terms as they are published. There is no interval to set and no scan cost. * **Scheduled scans:** X (Twitter), LinkedIn and Hacker News. SnitchFeed searches each platform for each term on a schedule you choose per listener, and each scan uses credits. ## Set the scan interval [#set-the-scan-interval] Open a listener and select **Scan Intervals**. There is one card per scheduled platform (**Twitter**, **LinkedIn** and **Hacker News**). A card is **Active** when at least one term has that platform turned on, or, for X and LinkedIn, when the listener tracks a profile on it. Pick an interval from the card's menu and click **Save** at the top of the editor. The interval applies to every term on this listener for that platform. | Interval | Label in the app | | ------------------ | ---------------- | | **Every hour** | Fastest | | **Every 2 hours** | Responsive | | **Every 4 hours** | Balanced | | **Every 6 hours** | Light | | **Every 12 hours** | Economy | | **Every 24 hours** | Minimal | Defaults for a new listener: | Platform | Default interval | | ----------- | ---------------- | | X | Every 4 hours | | LinkedIn | Every 12 hours | | Hacker News | Every 6 hours | Shorter intervals find mentions sooner. Longer intervals use fewer credits. ## What scans cost [#what-scans-cost] Each scan of one term on one platform costs: | Platform | Credits per scan | | ----------- | ------------------------------------------------------------------------------ | | X | 2 | | LinkedIn | 4, multiplied by the number of author job titles on the term (when it has any) | | Hacker News | 2 | [Tracked profiles](/docs/listeners/tracked-profiles) are fetched on the same X or LinkedIn interval: 2 credits per X profile and 4 per LinkedIn profile. Monthly estimate for one platform: ```text credits ≈ (24 ÷ scan interval in hours) × 30 × number of terms × credits per scan ``` **Worked example:** a listener has 2 terms. Both run on X every 4 hours, and both run on LinkedIn every 12 hours with no job title filter. * X: (24 ÷ 4) × 30 × 2 × 2 = **720** credits per month * LinkedIn: (24 ÷ 12) × 30 × 2 × 4 = **480** credits per month * Total for scans: **1,200** credits per month Each card shows the estimate for its platform, and the **Required credits/mo** box under the section list shows the total for the listener. These estimates cover scans only. AI scoring of each stored mention is charged on top. See [Credits and usage](/docs/credits). When your workspace runs out of credits, scans and real-time matching stop until credits are available again. See [Plans, billing and credit top-ups](/docs/workspace/plans-and-billing). ## Pause and resume a listener [#pause-and-resume-a-listener] Pause a listener to stop collecting without deleting it. Open the listener in **Listeners** and click **Paused** next to **Save**. Click **Active** to resume. The change applies right away, without saving the listener. These buttons appear on saved listeners only. While a listener is paused: * No scans run and no credits are used for its terms or tracked profiles. * Reddit and Bluesky posts published during the pause are not collected. * Its terms and tracked profiles still count toward your plan's keyword limit, and the listener still counts toward your listener limit. * Existing mentions stay in your feeds. When you resume, SnitchFeed checks each X, LinkedIn and Hacker News term and scans right away if its interval has passed since the last scan. Resuming does not start a historical backfill. ## Historical backfill [#historical-backfill] So a new listener does not start empty, SnitchFeed fetches recent posts for each term once. This runs when: * you create a listener (every term), * you add a term to a listener, * you change a term's query, **Whole word** or **Case sensitive** setting, or * you turn on Reddit, X, LinkedIn or Hacker News for an existing term (that platform only). What each platform fetches, per term: | Platform | What is fetched | Cost | | ----------- | --------------------------------------------------------------------------------- | ------------------ | | Reddit | Matching posts and comments from the last 14 days, up to 50 posts and 50 comments | No search cost | | X | One search of recent posts | 2 credits | | LinkedIn | Up to 2 pages of recent posts, for each author job title | 4 credits per page | | Hacker News | Matching posts, up to 200 | 2 credits | | Bluesky | No backfill | | Every backfilled post that is stored as a mention is scored and charged the normal AI scoring cost. If your workspace is out of credits, the backfill is skipped. Backfilled mentions appear in your feeds like any other mention. Feed notifications cover mentions that arrive after you set them up. See [Set up notifications for a feed](/docs/notifications/per-feed). ## Related [#related] * [Create a listener](/docs/listeners/create): where the Scan Intervals section sits in setup. * [Boolean search queries](/docs/listeners/boolean-search): what counts as a term. * [Credits and usage](/docs/credits): every action that costs credits. * [Supported platforms and how often they refresh](/docs/platforms): platform coverage. * [Plans, billing and credit top-ups](/docs/workspace/plans-and-billing): listener and keyword limits per plan.