ajx.

Make agent experience part of product design.

Design for agents.
See what
they see.

A usability toolkit for the interfaces agents actually use: your CLI, docs, SDK, and API. Inspect the journey and turn friction into a better design.

Open your first experiment ↗View the evidence ↓
Canvas / Agent experience100%
What to observe / Docs

Can the first example actually run?

Observe discoverability and the gap between documented and actual behavior.

Quick start → command → result
↗
Fictional sample ask

Make the quick-start export command executable.

Observe → verify → improveInspect every assumption
Fits the tools you already use
Codex·Claude Code·Kiro CLI·Agent Skills-compatible hosts

The workflow

Make the experience
something you can inspect.

Start with a single task. Grow into a matrix of harnesses and models once you know what you want to learn.

01

Give it a clear brief.

Choose a task, a success check, and the access the agent needs. Define the harness, model, and environment.

02

Watch a fresh attempt.

A new worker tries the product. AJX records its actions and keeps the artifacts available in the local console.

03

Turn evidence into asks.

Inspect verification and the visual journey. Follow each ask back to events before deciding what to change.

Try the output first

A useful report starts
with what to improve.

Switch between the fictional runs. Compare their asks and verified outcomes, then open the underlying evidence.

Demo Export CLIFictional sample
Codex · synthetic fixture3 product asks
First product ask

Make the quick-start export command executable.

The documented --out flag fails. Command help reveals --output, which works.

Verified outcome✓ Succeeded
Elapsed time1m 48s
Recorded calls*5 / 1 failed
*Partial telemetry · invented metricsOpen full report ↗

Your first attempt

Give your agent
a good starting point.

Open your project in a skill-capable agent and paste this prompt. It installs the native CLI and matching skill, checks setup, and helps you plan a small first run.

  1. 1
    Bring a project and your tools.

    Use macOS or Linux, with a supported worker CLI installed and signed in.

  2. 2
    Review the plan before running.

    Agree on the task, success check, access, and expected model costs.

  3. 3
    Read the results together.

    Open the report and visual journey. Inspect the first product ask and decide what to improve.

Ready to repeat it? Add matrix runs or ajx ci acceptance checks to your pipeline.

Paste this into your agent 
Install the latest stable AJX CLI and skill by following
https://raw.githubusercontent.com/royosherove/ajx/main/docs/AGENT-QUICKSTART.md
Then load the installed AJX skill.

I want to use AJX to test jq.
Starting point: https://jqlang.org/manual/
Task: Filter a sample JSON inventory to in-stock items and export their names and prices as CSV.

Prepare any tools, generated sample data, and local test server needed in a temporary workspace.
Use the AJX skill to run one trial with a fresh agent, verify the result, and show me the report and journey.
Try an example. No project needed.CLI + skill installed together
Prefer a terminal? Install the CLI + skill yourself. 
bash -o pipefail -c 'curl -fsSL https://github.com/royosherove/ajx/releases/latest/download/install.sh | bash' && export PATH="$HOME/.local/bin:$PATH"

macOS or Linux. The installer includes the skill. Your agent can read it directly; start a fresh session if the host needs one to discover /ajx. No GitHub login required.

Existing skill? Get the complete replacement command.

The Go CLI and Agent Skill install together.
No Go toolchain or GitHub login needed.