Hi everyone! I’m looking for an AI tool to generate creative event posters, ideally with an API for developers. I’ve tried general image generators like OpenAI, but the results often feel quite similar. I’m looking for something that can produce more original and varied poster designs. Any recommendations?

I've got an AI voice receptionist (Vapi + Vonage) set up for a therapy clinic client. Right now his business number (a normal personal mobile, not a VoIP line) forwards to the AI only on no-answer after 3 rings — works fine during business hours. What I actually want is for calls to skip his phone entirely and go straight to the AI number after his closing time (e.g. 8pm weekdays and 5pm weekends), without him having to manually toggle anything or rely on just not picking up. As far as I can tell, standard GSM forwarding codes (**21*, **61* etc.) are just on/off switches — no time-of-day logic. So it seems like true schedule-based routing would mean moving him onto some kind of virtual/VoIP business number instead of his existing personal mobile line. Has anyone actually solved this for a client who wants to keep their existing mobile number? Is there a cheap/simple way to get time-based forwarding without a full number port, or is "a proper VoIP business line with schedule rules" really the only real answer?

Context: I'm building an agent the screens candidates for recruitment companies on Vapi. The entire goal of the agent is to capture certain data about the candidate, so the entire call is the agent asking the candidate a bunch of questions Issue: Some questions that are listed in the prompt, the candidate already answers them indirectly or implies an answer from his speech, but the AI will ignore that and will keep on asking the question again Example: Agent: Can you tell me a little about your current situation and what has prompted you to consider a move? Candidate: I am an AI engineer at Google and I'm just exploring the job market. Agent: Thanks for sharing that. Are you currently employed at Google, and how actively are you looking to move now. So the agent asked the candidate whether they're employed at google after the candidate said he's an engineer at google I tried adding this piece of prompt: "## 3. The Question Criteria NEVER ask a question that the candidate has already gave you the answer for directly or implied, before asking a question, analyze the candidate's previous response and see whether the candidate has already gave you the info, if they gave you the info but you need further clarification in order to capture the required info, then ask for clarification" But the issue is still present more often than not

I’d really appreciate it if someone could share one. 🙏

What is the best model for now in OCTOBER 2026 to replace a character from a video as I am trying to replace myself on the video from a youtube video and also want to replace my own voice for that I was trying to use Higgsfiled genjutsu . and Also heard about the Kling O1 . But currently in Higgsfield it shows as Image model not the video generation model what is the best solution for this