It takes 2 mins to set up:
https://t.co/E6EfmblILx → create account
claude → settings → add custom integration
paste: https://t.co/ZYofecZC5C
comment Pikes MCP and I will send you 200 credits to try it out :)
claude is the hottest product in tech right now.
but it can't generate images — until today.
we just shipped pikes mcp to give claude a full creative studio.
generate product images.
remix products into lifestyle scenes.
Resize to any format.
animate stills into video.
under the hood, we have a suite of top image and video models, from nano banana pro (up to 4K), Ideogram, Flux, to kling, veo 3.1, seedance — auto-selected per task.
Consumer Startups’ @_leoluo sits down with our CEO & co-founder @JustinePalefsky to discuss the founding journey of Kindred, key lessons learned, and how the business evolved to where it is today.
Read here: https://t.co/AWfWRpUGJZ
.@bubblelab_ai supercharges your ops work in Slack. Deploy Pearl in one click, connect it to tools like Notion, Jira, Stripe, and let it run tasks and automations for your team directly in Slack.
Congrats on the launch, @Selinaliyy and @zhubzyz!
https://t.co/Zne13WV5Ic
.@_leoluo just published a deep dive on Duckbill that captures something most people miss about AI companies:
We're not trying to replace humans. We're orchestrating them.
A few thoughts on what he got right 🧵
Today we’re releasing Chroma 1.0
→ the world first open-source, end-to-end, real-time speech-to-speech model
→ with personalized voice cloning
Trained by FlashLabs.
Deployed on FlashAI👉 https://t.co/SeQyoJXKzt
An open research-grade alternative to the @OpenAI Realtime model.
Voice Test dubbing @elonmusk and @lexfridman: https://t.co/2lxvJD22Sl
🔥What’s real (evals and benchmarks attached):
⚡ <150ms TTFT (end-to-end)
🎙️ Native speech-to-speech (no ASR → LLM → TTS pipeline)
🧬 Few-second reference → high-fidelity voice cloning
📈 SIM = 0.817
→ +10.96% vs human baseline (0.73)
→ Best among open & closed baselines
🧠 Strong reasoning & dialogue with just 4B params (@Alibaba_Qwen 2.5-Omni-3B, Llama 3, and Mimi)
🔓 Fully open-source (code + weights)
With SGLang @lmsysorg enabled:
• 🧠 Thinker TTFT ↓ ~15%
• ⏱️ End-to-end TTFT ~135ms
• 🔊 RTF ≈ 0.47–0.51 ( >2× faster than real-time )
📘 SGLang Cookbook: https://t.co/pkw66tg8YM
📄 Paper + benchmarks: https://t.co/vfyYnO5rAg
🤗 Models: https://t.co/7jDWOD3kyk
💻 Inference code: https://t.co/EcJzhPBKmS
🔁 RT if you believe real-time AI should be open
💬 Reply if you’re building on Conversational Voice AI products
Just built my first tool using Claude Code. I’ve been wanting to build this one for a while...
It lets me polish up any text on any website, email client, or app with a simple keyboard shortcut, without having to copy and paste into ChatGPT (used to do that like 20 times a day)
Some of my favorite product images generated on Pikes AI lately.
We’ve cracked one of the toughest problems in AI product photography: keeping every product detail consistent.
If you’re running a brand and want to try our consistency model, shoot me a DM.
🎉 Excited to share that Gemini 3 Pro / Nano Banana Pro is now live on Pikes.
It’s easily one of the best image models out there, especially for lifestyle shots and more creative concepts. And yes — the outputs come in 4K.
If you want to try it out, head over to Pikes [dot] AI.
@TheSzef@TheSzef Tried it on my end as well. Let me know what you think.
It looks consistent with the hats on your website. Not sure which model or prompt set you used, but we have a consistency model that’s designed to preserve product details if you want to try that too.