Skip to content

10 curated AI tools for Coding & Software Agents

Coding agents complete code, review changes and run tests inside your editor or repository, helping developers ship software with less manual work.

More on this category

What these tools are

Coding agents work inside your editor, terminal or repository: reading existing code, proposing changes, running tests and opening pull requests. They are used by developers and teams shipping software continuously.

What they usually do

Completion and refactoring, implementing a described feature, tracing an error to its cause, generating tests, writing migration scripts and keeping changes consistent across files. Some run commands to verify their own work; others only suggest and leave execution to you.

Strengths and limits

They save time on boilerplate, repetitive edits and triage. The risk is confident edits in the wrong place, or changes that ignore project conventions. Larger codebases with real test suites suit them best.

How pricing works

Per-seat subscriptions are standard, with quotas measured in requests, completions or credits. Some bill per token, or let you bring your own API key.

How to choose
  • Check whether it understands the whole repository, not just the open file
  • Prefer tools that show diffs you can accept one by one
  • Look at sandboxing and permissions: what it can read, and whether it can reach the network
  • Confirm it fits your editor, code host and CI

Maintained by steven · About our listings →