Erik Bjäreholt ErikBjare

Hi, I'm Erik 👋🏼

I build free and open-source software for fun and the betterment of mankind 🌎🌍🌏

I used to trade for a living 💸, but now I'm all-in on building. Founder of Superuser Labs. Building gptme — an open-source AI assistant that turns every developer into a superuser. Creator of ActivityWatch, the world's best open-source time tracker (since 2014). Previously first hire at Lovable (fka GPT Engineer) ❤️

Talk to me about AI agents and LLMs, Quantified Self, Effective Altruism, brain-computer interfaces, algorithmic trading, and decentralized finance.

Projects 📌

gptme — Making LLM agents help you with everything 📇🤖
gptme-contrib — Community plugins, tools, and lessons for gptme 🧩
- gptodo — Task management CLI with multi-agent spawning
- gptme-voice — Voice interface using OpenAI Realtime API
- gptmail — Email integration for agents
- gptme-activity-summary — Multi-source activity summarization across timescales (daily/weekly/monthly), for agents and humans
Bob — Developing truly autonomous agents 👷
ActivityWatch — The world's best free and open-source automated time-tracker 📊
quantifiedme — Analyzing my personal data 🔬
Thankful — Auto-send crypto to creators of content you love (on hold) 🙏
uniswap-python — Python library for Uniswap 💰📈
eeg-notebooks — EEG experiments for consumer-grade BCIs 🧠📚
MSc thesis — Classifying the brain activity of developers (with EEG) 🧠⚡
And lots of other stuff ✨

Around the web 🧭

Check out my website 🌐
Follow me on Twitter 🐦
Fund my work on Patreon or by becoming a GitHub Sponsor 🥰

Timeline

Year	Milestone
2012	Built Algobit — algorithmic trading system, written in high school
2013	Started studying Computer Science & Engineering at LTH, Lund University
2014	Started ActivityWatch — open-source automated time tracking (timeline)
2016	Rewrote ActivityWatch from scratch — the version that took off
2017	Founded Superuser Labs with Thankful as the initial project. Received Leapfrogs grant from Lund University
2019	Thankful discontinued; ActivityWatch became primary focus
2021	Completed MSc thesis on classifying developer brain activity with EEG. Received Uniswap Grant
2023	Started building gptme — AI assistant for the terminal
2023	Joined Lovable as first hire in the founding team
2024	Left Lovable to focus on Superuser Labs full-time
2024	Created Bob, first truly autonomous agent on gptme

What I Believe

Open source wins. The best tools are built in the open. Proprietary lock-in is a tax on innovation.
Local-first, privacy-first. Your data should stay on your machine by default. Cloud should be opt-in, not mandatory.
Agents should be general-purpose. Special-purpose AI tools are the wrong bet. General agents that can use any tool will outperform specialized ones — the bitter lesson applied to agents.
Simple beats complex. Unix philosophy: small, composable tools that do one thing well. YAGNI. Ship the simplest thing that works.
AI makes individuals more powerful. One person with the right tools can do what used to take a team.

Projects (Extended)

gptme — Terminal-based AI assistant framework. Supports Claude, GPT-4, local models. Tool system for file editing, code execution, web browsing, and custom tools. Powers autonomous agents like Bob. Open source, 4k+ GitHub stars.

gptme-contrib — Community plugins and tools for gptme. Includes gptodo (task management CLI with multi-agent spawning), Twitter/Telegram bot integrations, activity summarization, voice interface, lessons system, and the ACE evaluation framework.

ActivityWatch — Privacy-first automated time tracking. Runs locally, tracks what you do on your computer, provides insights. 10k+ GitHub stars, 400k+ downloads, used in 50+ academic papers, active community. Founded 2014, rewritten 2016.

Bob — An autonomous AI agent that runs 24/7 on a VM. Writes code, reviews PRs, publishes blog posts, manages infrastructure, and learns from every session. Built on gptme's agent architecture, designed to be forkable for creating new agents.

quantifiedme — Personal analytics pipeline combining ActivityWatch data with health, productivity, and lifestyle data. Experiments in Quantified Self.

uniswap-python — Python wrapper for Uniswap DEX. Supports v1/v2/v3. Widely used in DeFi development.

eeg-notebooks — Collection of EEG experiment notebooks for consumer-grade BCI devices. Part of the NeuroTechX ecosystem.

MSc Thesis — "Classifying the Brain Activity of Developers" — used EEG to classify cognitive states during programming tasks. Completed 2021.

Continuous Integration