TachlisAI checks messages, answers, and images with its content classifiers, so conversations stay appropriate. These safeguards are always on, for every account and on every plan, and they can't be turned off.
What happens when a request is blocked?
TachlisAI doesn't answer and replies with a short message, such as "I can't help with that request." For some topics, it suggests you discuss the question with your rav or a mentor instead.
If you have an accountability partner, a note under the reply tells you they were notified.
What is the "Hashkafa filter applied" note?
On sensitive topics, TachlisAI answers according to the Hashkafa profile you chose. When it does, you may see Hashkafa filter applied under the answer. Select Change setting to review your profile. See Choosing your Hashkafa profile.
Do the safeguards apply everywhere?
Yes. They apply on the web, in the apps, on WhatsApp and SMS, to images you create or edit, and to every model, including Claude, Gemini, Grok, and Muse Spark.
What if I try to get around them?
TachlisAI recognizes attempts to bypass its safeguards. It replies "It looks like you're trying to bypass TachlisAI's safety features," and repeated attempts can block the account. See Locked or blocked accounts.
TachlisAI blocked something it shouldn't have
Classifiers sometimes make mistakes. Try rephrasing your question with more context. Then please tell us with the Thumbs down button or the flag (Report bug) under the reply, so we can improve. See Reporting a bug or a bad answer.
Something got through that shouldn't have
Please report it right away with the flag under the answer, and choose Unsafe content if you use thumbs down.