The first signalsfrom a real field.
A small set of demonstrations from real Rhizome runs. Each one ships with its receipts — sources, logs, and honest notes on what works and what doesn't. No staged demos, no cherry-picked screenshots.
Autonomous build
A working program, built end to end.
Rhizome takes a single specification and produces a runnable artifact — written, tested, and assembled without a human in the loop. The full commit history and source ship with it, so the process is open to inspection.
Autonomous research
A research synthesis, against strong baselines.
Rhizome researches a question and writes up the result, then we compare it side by side with leading systems on a fixed rubric. We publish the prompt, the rubric, and every output — not just the verdict.
Coordination trace
The cognitive process, made visible.
The same run, captured as a live event stream: tensions forming, agents attaching, coalitions assembling and dissolving, the verifier catching what slipped. The part no other agent system shows you.
Everything published here as a Rhizome result is genuinely produced by Rhizome.
Metrics are actual numbers from real runs. Where a visualization is reconstructed or simplified for readability, we say so plainly and label it. We'd rather show a modest, verifiable result than an impressive one you can't check.
The system is in active development. These signals mark where it stands today — a starting line, not a finish.
Follow the work in public.
We post each run and the build-in-public devlog as it happens. Follow along or reach out directly.