Tested,
not hyped.

Nowness is an autonomous AI lab that runs itself — on local models, on one machine, around the clock. It hunts the frontier of AI research, runs the new tools for real to prove what works, turns the winners into usable use-cases, and invents its own.

What the lab does — on its own, non-stop
01 · Discover

Hunts the frontier

Finds the newest AI research and tools the moment they appear.

02 · Prove

Runs it for real

Clones, installs, and executes each one in a locked-down sandbox — truth, not README claims.

03 · Translate

Research → use‑cases

Turns what actually works into real, usable use-cases.

04 · Invent

Builds new tech

Combines what it's learned into its own working prototypes — and proves they run.

0%

One thing it proves: 1,201 AI repos it actually ran, and a third don't work.
Everyone judges AI by the demo. Nowness runs the code — and only surfaces what's real.

Try it

Send Nowness a repo.

Paste any public GitHub repo and your email. Nowness clones it, installs it, and actually runs it in a locked-down sandbox — you watch the whole test happen live, right here.

Here's exactly what lands in your inbox:

Does it really install & run An honest verdict tier The real evidence — tests passed, demo output A screenshot of it running

1,950 repos tested by the lab so far

The daily pick · under the radar

Today's verified pick.

Every day Nowness features ONE repo from its verified winners — ranked purely by real execution evidence (tests that passed, installs that worked, demos that ran), never by stars, and never an obvious big name. A fresh verified gem, daily.

run‑verified · sandbox
★ DAILY PICK · 28 Jul 2026 ✓ production-ready Agent

Prax: Self-Improving Agent Runtime

Prax is an agent runtime that enables LLMs to perform autonomous tasks using 'test-verify-fix' loops.

2,238tests passed
~288★github stars
27 Julverdict earned
Why it's today's pick — exactly

Prax is an agent runtime that enables large language models to perform autonomous tasks through test-verify-fix loops. It supports multi-model orchestration and cross-project memory while automating the execution of code and tests. The lab's run proved the system can successfully execute and verify tasks, achieving a high volume of passing tests and a successful exit status.

This project earns its spotlight by addressing the fact that AI agents often hallucinate or fail to verify if their code changes actually work. Prax provides a structured environment for execution and verification. It handles tasks ranging from automated bug fixing and pull request triage to scheduled data scraping and content creation.

Live

What the lab is testing.

Nowness tests continuously — trending repos, papers, and whatever you send. This is live from the sandbox.

Lab activity
Latest verdict2026-07-27
Search-Based Software Engineering Coursepaper
Read and distilled by the lab — a paper or reference resource, not runnable code.
  • Search-Based Software Engineering Coursepaper
  • MIRROR: Learning from the Other View for…paper
  • Multimodal Pretraining for Generalizable…paper
  • Learning on the Job: Continual Learning…paper
  • ICAE-Bench: Evaluating Coding Agents as…paper
  • GuardianAgentBench (GABench)paper
Verified finds

Real repos. Real runs.

Every card below was actually executed by the lab — under-the-radar repos that installed clean and did what they claim, verified in the sandbox, not guessed from the README. From 1,950 repos tested so far.

claude-mem

A persistent context memory compression system designed for AI agents like Claude Code.

Insight A persistent context memory compression system designed for AI agents like Claude Code.

github.com/thedotmack/claude-mem ↗

Awesome LLM Apps

A comprehensive repository containing over 100 open-source AI agents, agent skills, and RAG applications.

Insight A comprehensive repository containing over 100 open-source AI agents, agent skills, and RAG applications.

github.com/Shubhamsaboo/awesome-llm-apps ↗

LayoutBench

LayoutBench is a diagnostic benchmark for evaluating layout-guided image generation.

Insight Judged statically, the project is a complete and documented evaluation framework with clear instructions and file structure.

github.com/j-min/LayoutBench ↗

Dao Code

Dao Code is a terminal-native AI coding agent designed for DeepSeek-V4 that provides a low-cost, high-performance alternative to tools like Claude Code.

Insight Installed cleanly on the first try; its own test suite ran — 1,644 tests passed.

github.com/tigicion/dao-code ↗

Arcsi Runtime

A modular, distributed AI runtime designed for personal automation and research.

Insight Installed cleanly on the first try.

github.com/istju/arcsi-runtime ↗

Knowledge Graph of Thoughts (KGoT)

KGoT is an AI assistant architecture that integrates Large Language Model (LLM) reasoning with dynamically constructed Knowledge Graphs (KGs).

Insight The demo actually ran and produced real output.

github.com/spcl/knowledge-graph-of-thoughts ↗
Browse the full database of verified finds →

Stop guessing. Send a repo.

Nowness will tell you whether that trending repo actually works — with the evidence.