2 comments

  • ShamilDibirov 23 hours ago
    Hi! It is cool and works pretty well! Did not you think to sell it to business as a replacement for hr? I am currently building an HR agent, now finished with cv screening functionality and wanted to implement phone call screening for candidates based on CV of applicant and job requirements, basically the same thing you did, but to sell it to hiring side.

    And if not a secret what model do you use? Is it gemini's multimodal real time API?

    • yomwolde 21 hours ago
      It started off as me just trying to solve a personal problem, but you're right, I've considered that direction. However, I'm staying focused on the candidate side since I still have a lot of ideas I want to explore there first.

      As for the LLM, I was using Claude 3.7 for a while but recently switched to GPT-4o.

      • stormfather 8 hours ago
        Are you using something like Pipecat or the VAPI API? How are you handling your voice->voice pipeline and your audio infra?
  • erikig 1 day ago
    Kudos! I just went through a "See it in action" demo and this is pretty great for job seekers. The UI works well to track the agent's and candidate's speech modulation.

    From the business side, how will you monetize given that you're catering to users that aren't the most likely to have disposable income?

    • yomwolde 1 day ago
      Thanks for trying it out and the kind words!

      Initially, SpeakFast was focused on general speaking skills, but pivoted when I found users were more willing to pay for targeted outcomes like company specific interviews—higher stakes, clear value, and immediate ROI