Augment Intent
Also known as: Intent by Augment, Intent, Augment Code Intent
Most AI coding tools are one-agent-at-a-time: you prompt, it responds, you prompt again. Augment Intent is designed for the coordination problem that surfaces when a task spans multiple services or repositories. You write a living spec (a document that describes the scope and requirements of the change), and Intent's Coordinator agent breaks it into discrete tasks. Implementor agents then run those tasks in parallel, each in its own isolated git worktree (a separate working copy of the repository), so they do not collide.
A Verifier agent checks completed work against the original spec before anything surfaces for human review. The living spec updates as the agents work, so the plan always reflects what was actually built. Intent also supports BYOA (Bring Your Own Agent), meaning you can use Claude Code, Codex, or OpenCode inside Intent's workspace without an additional license.
Intent launched in public beta in February 2026 and targets engineering teams working on complex, multi-service codebases where the bottleneck is coordination overhead rather than raw typing speed. The positioning is explicit: it frames itself as what comes after the IDE, for workflows where interactive back-and-forth in an editor is too slow and too manual to handle the scope of work.