Skip to content
@node-q

node-q

Coding agents, benchmarked. The team behind Harness Playground.
README.md 
███╗   ██╗ ██████╗ ██████╗ ███████╗       ██████╗
████╗  ██║██╔═══██╗██╔══██╗██╔════╝      ██╔═══██╗
██╔██╗ ██║██║   ██║██║  ██║█████╗  █████╗██║   ██║
██║╚██╗██║██║   ██║██║  ██║██╔══╝  ╚════╝██║▄▄ ██║
██║ ╚████║╚██████╔╝██████╔╝███████╗      ╚██████╔╝
╚═╝  ╚═══╝ ╚═════╝ ╚═════╝ ╚══════╝       ╚══▀▀═╝

╔════════════════════════════════════════════════╗
║  self-evolving AI, benchmarked  ·  nodeq.tech  ║
╚════════════════════════════════════════════════╝

node-q

Self-Evolving AI, Benchmarked for Trust.

As AI agents learn to improve themselves, progress needs proof. We build Harness Playground, where every result is pinned to three things: the harness and its version, the model, and the tests.

  • Prove it got better. Run each version of your harness on the same model and tasks, in the open, so anyone can check whether it improved.
  • Follow your harness over time. Upload your benchmark jobs. Each version gets its own row on a board, with a 95% interval, next to other harnesses. See the benchmarks.
  • Evaluate harnesses for your domain. Not everything is coding. If you don't know how to build a harness for your field, or don't want to hand your work to a third party, we can set up a private site on your own tasks, with its own database and domain, and none of it public. Talk to us.

hbench, our open source engine, runs Claude Code, Codex, Gemini CLI and 40+ more.

Open source

██╗  ██╗██████╗ ███████╗███╗   ██╗ ██████╗██╗  ██╗
██║  ██║██╔══██╗██╔════╝████╗  ██║██╔════╝██║  ██║
███████║██████╔╝█████╗  ██╔██╗ ██║██║     ███████║
██╔══██║██╔══██╗██╔══╝  ██║╚██╗██║██║     ██╔══██║
██║  ██║██████╔╝███████╗██║ ╚████║╚██████╗██║  ██║
╚═╝  ╚═╝╚═════╝ ╚══════╝╚═╝  ╚═══╝ ╚═════╝╚═╝  ╚═╝
Repository What it is
hbench The benchmark engine behind Harness Playground. Run coding agents on benchmark tasks from your terminal.
hbench-action A GitHub Action for hbench. Find out which coding agent is best for your code, in CI.
hbench-action-example An example repository: three tasks and the workflows that run coding agents on them.
hbench-datasets Benchmark datasets for hbench, published as signed OCI artifacts.

Quick start

You need Docker, on Linux or macOS.

curl -fsSL https://nodeq.tech/install.sh | sh
export ANTHROPIC_API_KEY=sk-ant-...
hbench run

Then upload the job folder to Your bench on nodeq.tech. It stays private until you publish it.

Links

Popular repositories Loading

  1. hbench hbench Public

    Benchmark runner for coding-agent harnesses, the engine behind Harness Playground

    Rust 17 1

  2. hbench-releases hbench-releases Public

    Release binaries of hbench, the benchmark runner for Harness Playground (nodeq.tech)

  3. hbench-datasets hbench-datasets Public

    Benchmark datasets for hbench, published as signed OCI artifacts

    Python

  4. hbench-action hbench-action Public

    Which coding agent is best for your code? Run Claude Code, Codex, Gemini CLI and 40+ more on your tasks in CI and find out.

    Python

  5. hbench-action-example hbench-action-example Public

    Example repository for node-q/hbench-action: three benchmark tasks and the workflows that check them and run coding agents on them

    Shell 1

  6. .github .github Public

    node-q organization profile

Repositories

Showing 6 of 6 repositories

People

This organization has no public members. You must be a member to see who’s a part of this organization.

Top languages

Loading…

Most used topics

Loading…