Futurelab Studio builds Ora, htmlctl, and a broader body of frontier software across local AI, research workflows, Telegram agent interfaces, and macOS tooling. The goal is not maximal surface area. It is software that feels deliberate, fast, and genuinely useful.
Current Signals
Futurelab is strongest when the homepage does not just describe a thesis, but shows what is shipping, what people are reacting to, and where the most interesting momentum is gathering.
What People Say
"dude I installed this and in like a day I already got it working better than my current openclaw... Way less overhead than openclaw. And any feature I miss I just ask pi"
Pablo on TelePi →"Connect Pi with Telegram using telepi. It gives you the gist of Claw without all of the fluff & complexity... I can already see big productivity improvement on todo list."
Filip Zivanovic on TelePi →Current Focus
TelePi and TeleCodex are where the fastest current iteration is happening: mobile supervision, ASR, attachments, profile switching, and real CLI handback.
Telegram Agent Cluster
TelePi and TeleCodex are converging on a workflow that feels more useful than it sounds at first glance: run coding agents from your phone with voice notes, screenshots, inline controls, and deliberate handback to the terminal when the work needs a real keyboard again.
People are not just reading this cluster. They are bending it into their own GTD and remote-coding workflows.
The phone becomes useful when the interface respects what phones are actually good at instead of mimicking a desktop IDE.
Ora
Ora remains the clearest Futurelab expression of voice, local models, native Mac actions, and privacy-preserving AI as a real product instead of a demo stack.
Ora is where Futurelab works out what privacy, speed, and native interaction should feel like on end-user software.
htmlctl
htmlctl is where Futurelab sharpens a different instinct: fewer hidden layers, explicit release state, and publishing infrastructure that stays readable to both humans and agents.
htmlctl is where Futurelab works out how release systems should behave when agents are first-class operators too.
Frontier Work
These projects are not random extras. They are where Futurelab tests interface ideas, research workflows, and local AI patterns before or alongside the flagship products.
Agent Interfaces
TelePi and TeleCodex are the clearest examples: voice notes, screenshots, profile switching, and real handback between phone and terminal.
Research & Documents
Scientific-paper tooling and OCR experiments point toward a larger Futurelab interest in reading, extraction, and structured knowledge work.
Local Mac Tools
MacTalk and OllamaCompanion show the same studio instinct in a different register: keep the workflow native, local, and shorter than the usual stack.
Futurelab is most interesting as a studio when the products, experiments, and essays feel like parts of one worldview rather than separate launches.
See the broader GitHub surface →Latest Writing
The most useful Futurelab pages are not only product pages. The supporting essays explain how the tools fit real workflows, where they differ, and why they were built the way they were.
Why Telegram works as a supervision layer once the architecture work is done and the loop becomes dispatch, review, and handoff.
ComparisonA practical breakdown of context model, mobile UX, ASR, launch profiles, attachments, and where each bridge is strongest.
OraThe privacy and latency case for local speech pipelines, and why that boundary matters more than voice features alone.
htmlctlWhy preview confidence matters more when the artifact actually stays the same between review and production.
Get in touch
Product questions, collaborations, and thoughtful intros are welcome. If you mostly want release notes and architecture notes, the newsletter is the cleanest way to keep up without chasing individual posts.