AXI: Agent eXperience Interface

10 design principles for building agent-ergonomic apps.

AI agents interact with external services through two dominant paradigms today: CLIs which were originally built for humans, and structured tool protocols like MCP. Both impose significant overhead.

AXI is a new paradigm - agent-native CLI tools built from 10 design principles that treat token budget as a first-class constraint.

Results

Browser Benchmark

Evaluated across 490 runs (14 tasks × 7 conditions × 5 repeats) using Claude Sonnet 4.6:

GitHub Benchmark

Evaluated across 425 runs (17 tasks × 5 conditions × 5 repeats) using Claude Sonnet 4.6:

Claude Sonnet 4.6 is the model the published runs used, not a limit of the harness. Both benchmarks take --model, which is passed straight through to the agent CLI, so any Claude model works - for example --model claude-opus-4-6. Results for other models are simply not published here, and they may respond differently to output format.

Quick Start

Reference AXI implementations:

• gh-axi — GitHub operations
• chrome-devtools-axi — Browser automation

See the full AXI Catalog below for all official and community AXIs.

npm install -g gh-axi
npm install -g chrome-devtools-axi

Add to your CLAUDE.md or AGENTS.md:

Use `gh-axi` for GitHub and `chrome-devtools-axi` for browser automation.

The 10 Principles

These principles define what makes a CLI tool "an AXI". The table below is generated from principles.yaml; the full specification of each principle lives in the AXI skill.

AXI Catalog

The catalog tables below are generated from catalog.yaml - see CONTRIBUTING.md to add your AXI.

Official

Reference implementations maintained by the AXI project, validating the principles across different domains:

Community

AXIs built and maintained by the community:

Built an AXI? Follow the contributor workflow to add it to this list.

Build Your Own AXI

Install the AXI skill to get the design guidelines and scaffolding for building an AXI-compliant CLI:

npx skills add kunchenguid/axi

This installs the AXI skill - a detailed guide with examples for each principle that your coding agent can reference while building. For your own AXI, expose an explicit setup command for session hooks as the primary integration, then ship an installable Agent Skill as a lower-overhead secondary path. Users only need one path, but hooks and skills complement each other when both are available.

Development

Browser Benchmark

The browser benchmark harness lives in bench-browser/. It compares browser automation tools across 16 browsing tasks.

pnpm install

# Run a single condition × task
pnpm --dir bench-browser run bench -- run --condition chrome-devtools-axi --task read_static_page

# Run the full matrix
pnpm --dir bench-browser run bench -- matrix --repeat 5

# Generate summary report
pnpm --dir bench-browser run bench -- report

# Render the social video
pnpm --dir bench-browser run render:social

The HyperFrames composition for the social asset lives in bench-browser/social/. Edit social/index.html for the animation and render docs/social/rendered/race.mp4 with pnpm --dir bench-browser run render:social.

Published results (490 runs): bench-browser/published-results/report.md

GitHub Benchmark

The GitHub benchmark harness lives in bench-github/. It runs agent tasks across different interface conditions and grades results with an LLM judge.

pnpm install

# Run a single condition × task
pnpm --dir bench-github run bench -- run --condition axi --task merged_pr_ci_audit --repeat 5 --agent claude

# Run the full matrix
pnpm --dir bench-github run bench -- matrix --repeat 5 --agent claude

# Generate summary report
pnpm --dir bench-github run bench -- report

Published results (425 runs): bench-github/published-results/STUDY.md

Contributing

Contributions targeting main must be submitted through the contributor workflow, which uses no-mistakes and guards release-please-generated files from hand edits.

Links

• Website
• AXI Skill definition
• Browser benchmark study
• GitHub benchmark study