Which tools are built for AI agents?
In the agentic era, the tools worth using are the ones an AI can drive for you. We score each one by how it can be controlled, from a command line an agent can type into, to a plug-in standard, to a public API, down to needing a human to click around. Then compare the big options and learn the Claude Code features that make all of it run on its own.
Find your starting point
Two taps for a path that fits you. No sign-up.
Want a tailored stack and a readiness score for your exact situation?
Get guided setupHow we score
A higher score means an AI agent can run the tool with less human help. We weight the surfaces by how directly an agent can use them.
A command line the agent can type commands into. The most direct, scriptable control.
A first-party connector built on the Model Context Protocol, the shared standard for giving agents tools.
A public programming interface. Powerful, but the agent (or you) has to wire it up.
No real interface, so an agent has to click through the website like a person. Brittle and slow.
The scoreboard
Sort, filter, and search the tools. Tap any tool to see exactly what its command line, connector, and API offer.
Showing 24 of 24 tools, ranked by automation surface (CLI counts most, then official MCP, then API).
Compare the big options
Head-to-head breakdowns of the agents, editors, and ways to drive them from your phone.
Claude Code vs OpenAI Codex: Which AI Coding Agent Should You Use?
Cursor vs Claude Code: Which AI Coding Tool is Better?
How to Control AI Coding Agents from Your Phone
Terminals for the Agentic Era
Learn the internals
The Claude Code features that let work run on a schedule, in the cloud, and without you watching. Plain-language deep dives.
Scheduling and Routines
Three tiers of scheduled automation: local loops, cloud routines that run with your laptop closed, and Desktop scheduled tasks.
Read the deep diveEffort Levels and Model Configuration
Control how deeply Claude Code reasons on each step, from quick low effort responses to deep xhigh analysis, plus fast mode and the opusplan model alias.
Read the deep diveDynamic Workflows
JavaScript scripts that orchestrate up to 1,000 background agents for large scale tasks, keeping results out of the context window.
Read the deep diveHeadless and Automated Execution
Run Claude Code non-interactively for scripting, cron jobs, and programmatic integrations using print mode, the Agent SDK, and permission controls.
Read the deep dive