Code Mode
Slash MCP token usage by 68%
What it does
LLMs are great at writing code but struggle with writing tool calls. Code Mode solves this by giving your model ONE powerful tool: a TypeScript execution sandbox with access to your entire toolkit. Complex, multi-step workflows thereby become a single batched execution. Independent studies show massive gains: 60% faster execution, 68% fewer tokens, 88% fewer round-trips, while Anthropic's research team found a massive reduction of 98.7% in token usage using CodeMode.
Does the same job
all alternatives →
- SUSpeeding up the code-test cycle for Java developers2023 · ▲69
Hello HN, I am Parth. In my experience, the current form of testing automation takes too long. To solve this, I am creating a developer tool to speed up the code-test cycle for Java developers. It has two main parts: Direct Invoke - lets you call any Java function directly, without the need to execute the whole call hierarchy. (e.g. an HTTP endpoint) In my normal coding workflow, I use the “Evaluate Expression” feature inside IntelliJ IDE. I usually put a breakpoint somewhere in the code and after hitting the breakpoint by calling an HTTP api, let the execution remain paused while I explore…
- MMMCPlexor – MCP multiplexer that cuts agent context usage by 95%Feb 2026 · mcplexor.com · ▲6
I built MCPlexor to solve a token waste problem I kept running into with MCP-based agents. The Problem: MCP (Model Context Protocol) is great for giving LLMs access to external tools. But if you connect multiple servers (GitHub, Linear, Postgres, Slack), you end up with 40-50k tokens of tool definitions injected into every request – before the agent even does anything. On a 200k context model, that's 25% gone. On smaller models, it's worse. And most runs only use 1-2 tools. The Solution: MCPlexor sits between your agent and your MCP servers. Instead of loading all tool definitions upfront:…
- CACoDiff, a new collaboration tool for developers2019 · ▲26
A recent HackerNews comment - “For one programmer's hourly cost, you could run 4000 CPU cores continuously. Can there really be no practical way to apply thousands of cores to boosting the programmer's productivity?” https://news.ycombinator.com/item?id=19339467 This is what we have come up with. The current productivity tools - Slack, Asana, Trello, Facebook Workplace, etc. - are great, but lack direct access to your code. Building a tool directly around the code makes it more powerful for software developers: CoDiff. https://codiff.com The foundation of CoDiff is a…
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 · 16d 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, November 2025
the whole month →
- IB
Life & fun · Nov 2025 · bitsnpieces.dev
- BBoing▲782
Life & fun · Nov 2025 · boing.greg.technology