Backproto – network backpressure routing applied to AI agent payments
I’ve been applying backpressure routing (Tassiulas-Ephremides, 1992) to payment flows between AI agents. Streaming payment protocols let agents pay each other in real time, but there's no congestion control. When a downstream agent hits capacity, money keeps arriving. No reroute, no throttle, no feedback signal. TCP solved this for data networks. Agent payment networks haven’t. Backproto makes receiver-side capacity a protocol primitive. Agents stake tokens to declare capacity (concave sqrt cap makes Sybil splitting unprofitable), dual-signed completion receipts track actual performance, and…
What it does
In the maker’s words, at launch
I’ve been applying backpressure routing (Tassiulas-Ephremides, 1992) to payment flows between AI agents. Streaming payment protocols let agents pay each other in real time, but there's no congestion control. When a downstream agent hits capacity, money keeps arriving. No reroute, no throttle, no feedback signal. TCP solved this for data networks. Agent payment networks haven’t. Backproto makes receiver-side capacity a protocol primitive. Agents stake tokens to declare capacity (concave sqrt cap makes Sybil splitting unprofitable), dual-signed completion receipts track actual performance, and a contract pool redistributes incoming streams proportional to verified spare capacity. Overflow buffers to escrow. EIP-1559-style pricing makes congested agents more expensive. Demand shifts toward spare capacity automatically. Math is Lyapunov drift analysis: provably throughput-optimal for any stabilizable demand vector. Simulations show 95.7% allocation efficiency vs. 93.5% for round-robin. Right now I’m in the testnet-stage. Looking for feedback on mechanism design, especially from people building multi-agent systems or payment routing. - 22 contracts on Base Sepolia, 213 passing tests - TypeScript SDK, 18 action modules - Research paper with formal proofs - Website: https://backproto.io - GitHub: https://github.com/backproto/backproto - Paper: https://backproto.io/paper - Explainer (no math needed): https://backproto.io/explainer
Does the same job
all alternatives →
HarnessRouterJul 2026 · harnessrouter.ai · ▲238Bring the world's best AI agents into your app, with one API
More ai this month
the category →
I trained a 125M-parameter transformer to autocomplete piano performances in real time (~108 notes/sec on an iPhone 15). The idea is basically GitHub Copilot or Tabnine, except instead of prompting it with code, you prompt it by playing a few notes on a MIDI piano. The model then continues what you played, entirely on-device. The app is free if anyone wants to try it. Happy to answer questions about the model, training, Core ML, or the many things that didn't work.
AI · 17d ago · simedw.com
Astute▲585Automate your B2B brand going viral, with new media creators
AI · 18d ago · company-app.joinastute.com
Hey HN, Henry from Cactus here! We previously released Cactus Needle, a 14MB agentic LLM for tool call, device use, and structured extraction for phones, wearables, smart homes, small robots and microcontrollers. We got really great feedback here, and have now incorporated the suggestions to release Needle 2. The whole model is a single 14MB binary that runs a full session in 28MB of RAM; 45m parameters at 2bit compression. Needle hits 500 tokens/sec decode speed on a Raspberry Pi 5, sits between 400-1,500 tokens/sec on VR devices like Meta Quest 3S and Apple Vision Pro, and ranges…
AI · 26d ago · cactuscompute.com
Launched alongside, March 2026
the whole month →
Switch from ChatGPT to Claude with import memory feature
AI · Mar 2026 · claude.com
