← Puter stuff
Open on Instagram ↗

Comments (8)

  • @anshif_kolka 2026-03-26

    The cost of gemini embedding 2 model for video is 0.00079 per frame. So considering the least 24fps video one hour of footage will cost you around 68 dollars per hour of footage; if fed directly.

  • @dylan_caponi 2026-03-25

    So it’s super expensive because you’re running all the video through Gemini?

  • @accidentalanthropologist 2026-03-25

    Sounds token heavy

  • @kanishqvijay 2026-03-25

    Is it just a RAG with chuncks of video?

  • @suk_adiw 2026-03-25

    This is really useful, but also looks like its token heavy, if we can use local models that would be fine if we can run it locally that is

  • @jeremias.em 2026-03-26

    Amazing but hooooly... Most be extremely heavy

  • @janis_88527 2026-03-25

    Why wirh gemini why not use on device Lllm what a waste of tokens

  • @aman.nnd0 2026-03-25

    This is cool and easy, in my case , best feature is that, you can talk to telegram bot, and it will provide you the exact length file whenever you ask for it.. . Check out @inchtube youtube ,,videos there are made exactly using this concept.

Content ideas (0)

No ideas generated yet. Run /instagram-sync ideate from Claude Code to create some.