I threw away my analytics dashboard and replaced it with 42 MCP tools
Back in October, I built a web analytics tool. It had two distinguishing features: it was cheap, and it had a weekly AI email summary so you wouldn't have to check your dashboards. Turned out that cheap made no difference, but the few users I had really liked the email digest. The world moves fast these days, and suddenly I found myself experimenting with agents, and doing all my work in Claude Code, so I decided to throw out everything and rebuild the tool as lodd.dev, a headless web analytics tool for agents. 42 MCP tools and a full API that lets agents call and act on analytics data as…
What it does
In the maker’s words, at launch
Back in October, I built a web analytics tool. It had two distinguishing features: it was cheap, and it had a weekly AI email summary so you wouldn't have to check your dashboards. Turned out that cheap made no difference, but the few users I had really liked the email digest. The world moves fast these days, and suddenly I found myself experimenting with agents, and doing all my work in Claude Code, so I decided to throw out everything and rebuild the tool as lodd.dev, a headless web analytics tool for agents. 42 MCP tools and a full API that lets agents call and act on analytics data as part of their processes. I've tried to optimise it for agent usage, so efficient responses (a snapshot is 60ish tokens), an llms.txt to guide the agent to handle setup, read only for simplified auth, and both a hosted oAuth (for desktop and mobile apps) and stdio version using an API key for terminal use. Human in the loop authentication using OTP in an email. Explicitly asking the agent to use it works really well, and I particularly like combining with other context sources like GSC or commit history, but it's been a little bit of a challenge having the agent call it unprompted. A few lines at the top of a Claude.md file has made a difference though. I'd love to hear what value, if any, you see from being able to get analytics data in the same conversation as your code (or other automatic workflows)?
Does the same job
all alternatives →
- ALA lightweight open-source web analytics for webdevs2023 · github.com · ▲7
I wanted to share something I've been working on recently: Statum, a handy web analytics tool. https://github.com/extractumio/statum The journey began when I found myself frequently needing simple web analytics for my projects. I tried Google Analytics, especially GA4, and realized it was quite complex and, at times, not very accurate, especially when I wanted to view stats for the current day or recent hours. Then I tested a few fancy startup solution but ended up with way too expensive plans the expect me to subscribe (I'm not that rich to pay $99/month for every…
- TSThe simplest tiny analytics tool – storywise2023 · github.com · ▲9
I've built a little prototype a long while back, to help me keep track of visitors on my website. I didn't like google analytics for obvious privacy reasons, and I didn't like the existing self-hosted options because of how complicated they are. I always wanted a simple, basic tool, similar to plausible, but even simpler. So a few weeks ago I rebuilt my prototype into a reusable tool. And I made it open source. I built a website, a documentation, and I'd be excited for someone else to try it too. Who knows, maybe if enough people like the idea, I might even make it into a SaaS to save…
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 · 27d ago · cactuscompute.com
Launched alongside, May 2026
the whole month →
Parallel agents, diff reviewer, and multi-model comparisons
Dev tools · May 2026 · kilo.ai
- NW
Hey HN, Henry here from Cactus. We open-sourced Needle, a 26M parameter function-calling (tool use) model. It runs at 6000 tok/s prefill and 1200 tok/s decode on consumer devices. We were always frustrated by the little effort made towards building agentic models that run on budget phones, so we conducted investigations that led to an observation: agentic experiences are built upon tool calling, and massive models are overkill for it. Tool calling is fundamentally retrieval-and-assembly (match query to tool name, extract argument values, emit JSON), not reasoning. Cross-attention…
Life & fun · May 2026 · github.com
- FM
Dev tools · May 2026 · github.com