Open Source Radar — August 15: agents get browsers & control planes

Share
Open Source Radar — August 15: agents get browsers & control planes

The coding-agent parade paused today — none of this week's workspace bets repeat, so the board is fresh. The signal is the layer around agents: a browser built to share with yours, a control plane for work that runs for days, and DeepSeek's official playbook for wiring its models into every agent tool. Three projects worth your attention.

ego-lite (JavaScript, ~10.6k stars) — A macOS browser built from the ground up for agent automation: your agents run browser tasks in their own Spaces while your tabs stay yours, and they can reach your real logins without you handing over a session. The pitch is that it fixes what browser-use-style automation frameworks get wrong — no separate browser to drive, no logins that break mid-task, fewer tokens per task — and it installs its browser skill into whatever agent you already run, Codex or Claude Code. Four months after launch it's on today's daily board at 10.6k stars, with Windows and Linux on the roadmap. If your agents keep flailing at logged-in web work, this is the most direct fix on the board.


LoopX (Python, ~4.8k stars) — An open, provider-neutral control plane for long-horizon agent work that sits on top of Codex, Claude Code, Cursor, or your own harness instead of replacing it. It keeps the durable state — objectives, gates, typed todos, evidence, spend quota — outside the agent's chat memory, so work survives restarts, hands off cleanly between agents, and a human gets asked a concrete question when judgment is needed. That's the gap it names: an agent can finish a task in one session, but work that runs for days needs governance, and chat memory plus a timer isn't governance. Local-first and MIT-licensed, it's the most interesting attempt yet at making long-running agents reviewable instead of a black box you babysit.


awesome-deepseek-agent (~5.8k stars) — DeepSeek's official curated list of guides for running DeepSeek-V4-Pro and V4-Flash inside 20-plus agent tools — Claude Code, Codex, Cline, OpenCode, Qwen Code, Hermes, and a dozen more. Each guide walks through installation, configuration, and first run, which makes this the canonical map for teams that want DeepSeek's models as their agent backend without fighting scattered docs. It's a catalog, not a runtime — but it's also the clearest sign yet that DeepSeek is consolidating its ecosystem playbook, and it landed on today's board at 5.8k stars. If you're shopping for a model backend for your agents, this is the list to start from.

Worth watching this week: all three point the same way — agents are getting a proper operating layer, from shared browsers to durable state to official integration playbooks.

Would you hand your logged-in browser to an agent, or keep it sandboxed? Tell us in the comments.

Sources: ego-lite (GitHub) · LoopX (GitHub) · Awesome DeepSeek Agent (GitHub)

Read more

Akhetonics says its all-optical CPU reaches a customer in 2026

Akhetonics says its all-optical CPU reaches a customer in 2026

Light-based computing keeps promising more than it delivers — but one Munich startup has just put a date on its bet, and the interview laying it out is doing the rounds on Hacker News this week. Akhetonics says it will deploy its first commercial machine with a major customer by the end of 2026, with several more planned for 2027. The company, founded by Michael Kissner and Leonardo Del Bino, is building a computer where data enters as light, is switched as light, and circulates through memory

The Week in AI — October 5–11, 2026

The Week in AI — October 5–11, 2026

Every big claim this week turned out to rest on fine print more interesting than the headline: revenue only the company reporting it can define, safety tests sandboxed while the product keeps the web, and a Pentagon phase-out nobody would confirm until reporters kept asking. The week's top 5 1. OpenAI's revenue was $20 billion below the numbers everyone quoted — and the gap was definitional. The Financial Times reported Thursday that OpenAI's annualized revenue runs roughly $20 billion unde

Drone strike shuts a third Yandex data center, taking YandexGPT offline

Drone strike shuts a third Yandex data center, taking YandexGPT offline

Russia's largest tech company is learning what the AI era's infrastructure war looks like from the receiving end — three data centers in four days, and with them much of the cloud layer Russian businesses run on. A Ukrainian drone strike knocked out Yandex's data center in Vladimir early Sunday morning, the third of the company's facilities hit since October 8. The site — reported at roughly 50 MW and designed for about 2,880 server racks — stopped operating completely after the attack, Yandex

Agent teams cost up to 5x more, barely score higher

Agent teams cost up to 5x more, barely score higher

The multi-agent hype train hit a benchmark this weekend — and the grid and the trucking regulators had quiet weeks of their own. Vals AI put agent teams head-to-head with single agents on its Vibe Code Bench, and the teams cost between 1.8 and 5.1 times more for almost no extra quality. The evals company ran GPT-6 Sol and Claude Opus 5.5 solo and in teams across 50 apps at two reasoning efforts; out of four comparisons, only one was statistically significant — Sol at medium effort, where the t