We built Dropstone because we were tired of starting from zero every time we opened an AI coding tool.
Dropstone learns, remembers, and evolves with your projects building a persistent understanding of your codebase, architecture, and workflow. It s designed to grow with you, not reset after every session.
We ve just launched it on Product Hunt and would love your thoughts on how memory should shape the next generation of developer tools. Your feedback will help us refine what s coming next.
Dropstone 1.5 is our monthly re-baseline release. We test the top AI coding models every month and rebuild the runtime on whichever wins.
This month: DeepSeek V4 Flash, V4 Pro, and Moonshot Kimi K2.6 all hosted in the US, nothing stored on our side.
$15/month gets you about 450 deep coding sessions a week. Roughly twice what Claude Code Pro delivers at $20.
Full technical report with the cost math and benchmarks: blankline.org/research/dropstone-1-5
Dropstone is the first multiplayer AI workspace. v3.0.5 adds Share Chat: send a link to code with humans & agents in real-time. Features infinite context (D3 Engine), persistent memory & background swarms. Built on original research, not a wrapper.
Dropstone v3 breaks the "Linearity Barrier" in AI coding. Powered by the proprietary D3 Engine (Dynamic Distillation & Deployment), it replaces linear token prediction with a Recursive Swarm Architecture. It simulates 10,000+ divergent timelines to prune errors before they happen. Features Horizon Mode for architectural planning and Semantic Entropy Tracking for real-time hallucination defense.
Product Hunt is obsessed with "Fast" today. How fast can it generate a website? How fast can it write a SQL query?
But ask any founder or senior engineer: What is the most expensive thing in software? It isn't writing code. It's fixing code.
We realized the industry is optimizing for the wrong metric. Current AI tools (Chatbots) are like hyper-active interns. They type at 1,000 words per minute, but they make mistakes that take days to debug. They prioritize Output over Outcome.
If you just glance at the homepage today, you d think the biggest innovation in tech is a "Year in Code" wrapper.
At #3: GitHub Wrapped. (Vanity metrics).
At #1: A generalist workflow tool.
But sitting quietly at #5 is arguably the most significant architectural shift in AI coding we've seen this year.
I ve been digging into the "Launch" list, and Dropstone is the only tool that isn't optimizing for "Velocity" (Speed). They are optimizing for Physics.
We ve officially benchmarked Dropstone against the AGCI Benchmark a framework designed to evaluate adaptive and long-term cognitive intelligence in artificial systems.
Unlike traditional reasoning tests, AGCI measures how well systems learn, remember, and reason across sessions essentially assessing continuity, adaptability, and persistent understanding.
Dropstone: World's first generational recursive self-improvement AI for autonomous programming. Desktop application with MCP Server, Computer Use API, unlimited Ollama models, Claude Code integration, and Agent Mode.
