KINETK - The intelligence graph for the social web.
by•
KINETK turns the social web into a real-time IP Graph of creators, communities, and narratives — accessible through the KINETK API and MCP for AI agents, with continuous enrichment and provenance.
Replies
Hi Product Hunt. I'm Mariale, co-founder and CTO of KINETK.
Rather than explain the API, we built something you can poke: growth-demo.kinetk.ai.
Describe your product in one sentence. A $400 sleep-tracking ring for anxious millennials, say, or a weighted blanket for hot sleepers. Hit build, and an agent returns a full launch and growth plan, evidence-backed, grounded in live social signals from the KINETK graph.
The word that matters there is evidence-backed. Most tools would happily invent you a plan from thin air and a confident tone. This one reads the actual social web first: the creators who move that niche, the communities they live in, the narratives trending right now. Then it builds the plan from what is real, not from what the model half-remembers from last year.
Everything the demo does is just the public API and MCP doing their job. Whatever it builds for your product, you can build too. Docs at docs.kinetk.ai, free key at platform.kinetk.ai.
Try it on something absurd and tell me where the plan is scary-good and where it faceplants. That feedback beats any upvote.
Tried it with one of my YouTube ideas. I was impressed by how quickly it generated a structured launch plan with trend-backed recommendations. One thing I noticed, though it inferred parts of my concept that weren't actually there, so a bit more grounding in the original prompt would make it even stronger. Really interesting product. Congrats on the launch!
@aziz_daiya Hi glad you liked it. Thank you for your feedback, we want to make sure we are grounded and informative. Is there anything else you would like to see the API provide?
Thanks! One thing I'd find useful is a confidence indicator for each recommendation.
For example, showing whether a suggestion is based on the prompt itself, current trend data, or an inference made by the model. That would make it easier to distinguish between evidence-backed recommendations and AI assumptions, especially when planning content.
@aziz_daiya thanks for the feedback and suggestions. Very good point, as that is very important for the responses to truly impact users decisions. All the data returned in the api is backed by actual content records. Are you referring to adding a confidence indicator in the narratives response or insight response? Or what extra fields would you like to see in the json payload to give you that confidence?
@thinkmariale Thanks for the clarification!
To be honest, I can't confidently suggest what the JSON payload should look like—that's outside my area of expertise.
My perspective is as a Product Designer and end user of the product. When I tested it with a YouTube channel focused on toddler content, one recommendation suggested creating FIFA football videos. While that may have been based on trend data, it didn't align with the audience or niche I had provided.
From a UX perspective, I'd love the responses to better communicate why a recommendation is being made and how closely it matches the context of the original prompt. That extra transparency would make it much easier to trust and act on the recommendations.
Been digging into KINETK for a bit and the live graph view is wild, you can actually trace how a niche community formed around a meme through linked clips and posts. The API response times felt snappy too, even on some pretty gnarly semantic queries.
@nevzatakal1e4q Love that you are throwing gnarly semantic queries at it. That is exactly what we want the community to do. Try it, test it, break it. The weird edge cases are where we learn the most, so if you find one that trips it up, send it my way.
The graph view is genuinely cool for tracking how a niche topic spreads across creators in real time. Setup was painless and credits feel predictable compared to other APIs.
@ayferekiolrsu6 thats is great to hear. Here for any feedback you might have from using our API.
Mapped a few creator accounts and the graph view actually made connections I hadn't noticed before, especially around niche communities. The vector space feels genuinely useful rather than just buzzword infrastructure. Solid first look.
@keremlxy6c yeah, community and narrative intelligence is our whole focus. It is how you see what's really happening on the social web instead of one post at a time. Curious what niche you would point it at first.