Understand and verify interaction probabilities

A probability controls whether a workflow attempts a like, bookmark or follow. It is not a fixed quota or a platform-approved safety threshold. Small samples may differ substantially from the configured percentage.

Check the action switch first, then the probability. Review eligible items, previously processed records and outcomes after execution. Missing actions can result from filters, page errors or changed state, not only probability.

Do not test wording by increasing comment volume. Observe a small scope and stop on platform notices; randomness is not evidence of immunity from restrictions.

Configure from naming through comments

1. Describe the topic, region and purpose in the name/description so logs remain identifiable.
2. Browse recommendations without keywords, or add manually verified search terms.
3. Match the filter version to the site; older general/latest/hottest controls differ from newer sorting options.
4. Enable likes, bookmarks and follows separately and set probabilities, not guaranteed daily quotas.
5. Pro users choose custom comments, AI Style or AI Identity. Custom mode selects library text; AI Style currently offers eight styles.
6. Save, start manually and check page/actions/stopping before adding schedules.

Two misunderstandings about AI and schedules

The model does not automatically learn your private interaction habits merely through continued use. Check context and style. Weekday scheduling defines local start times and still needs Chrome and login; inspect stopping controls separately when automatic ending is needed.

Product reference

Product steps were checked against extension 3.5.3. See the related guide for instructions and pricing for access. Report behavior that differs through feedback, including the version and a redacted error.