Introducing hivememory
Shared reasoning memory for multi-agent systems.
Agents write structured claims with evidence + provenance, query what others already know before making LLM calls, and catch contradictions automatically.
pip install hivememory
ridxm.github.io/hivememory/
I built hivememory, a shared reasoning memory layer for multi-agent systems
agents write structured claims with evidence + provenance, query what others already know before researching, and catch contradictions automatically
benchmark results with 3 parallel agents:
- 56% of queries served from memory
- 17.5% token reduction
- quality equal or better
- conflicts identified
project page: ridxm.github.io/hivememory/
repo: github.com/ridxm/hivememory
inspired by @karpathy's LLM knowledge base. single-agent works, this makes it multi-agent.
pip install hivememory
claude-eyes is finally a native feature
xitter.cf/ridcursion/status/2013…
🏆 Excited to share I got 3rd place at the @GoogleDeepMind x @vercel Hackathon NYC
I built ANSR. An AI voice host that answers every restaurant phone call. Takes orders. Makes reservations. Responds in any language.
How it works:
→ Owner uploads a photo of their menu
→ Gemini 3.1 Pro Vision extracts every item, price, and dietary tag
→ @ElevenLabs powers the voice conversation
→ @GeminiApp extracts structured data from the conversation
→ @supabase Realtime pushes orders and reservations to a live dashboard
The entire pipeline (voice in, structured data out, dashboard updated) happens with zero human input.
Built with:
→ @vercel Nextjs
→ Gemini 3.1 Pro + Flash Lite
→ @ElevenLabs Conversational AI
→@supabase Postgres + Realtime
Huge thanks to @vercel, @GoogleDeepMind, and @cerebral_valley for putting this together :)
Introducing Clutter Bot
A robot that understands simple english commands and actually executes them.
Natural language → robot action. Made using @GeminiApp and @Vultr
Built with @danielwu28 for the @lablabai AI Meets Robotics hackathon!
went to sf for Physical AI hack!
we tried to teach a robot to do latte art. a really fun weekend, met awesome people and learned a ton!
Day 9: building 30 ai agents in 30 days!
pitch your startup idea to 5 AI investors.
each one has a different personality. they grill you, argue with each other, and decide if they're in or out.
would you pitch your idea?
I got tired of being claude's eyes
Every time @claudeai fixed something in my app, I had to check the browser, paste the console error, describe what I saw.
So I built claude-eyes - an MCP server that lets claude actually see what it's building
Here's the thing: the existing playwright and puppeteer MCP servers return file paths for screenshots. but claude can't see file paths. It's still blind.
claude-eyes returns actual image data using MCP image content blocks. Claude receives the screenshot as an image it can visually inspect and reason about.
before: claude writes code → asks me to check → i paste error → repeat 5x
now: claude writes code → takes screenshot → checks console → verifies fix → done
26 tools total, built for autonomous QA:
→ see: screenshot, visual_diff
→ read: get_console_logs, get_network_requests
→ interact: click, type, scroll, hover
→ verify: verify_no_errors, verify_element, verify_request_succeeded
stack: typescript, playwright, MCP SDK, pixelmatch for visual diffing, sharp for image processing
open source: github.com/ridxm/claude-eyes
would love feedback from anyone using claude code for frontend work!
day 8: building 30 ai agents in 30 days!
Meeting Prep Agent
→ give it a name, get a dossier
searches across google, news, github, youtube and more - then synthesizes who they are, recent work, and talking points
prep for any meeting in <30 seconds!
#buildinpublic