SynaraQuickstart
Go from download to your first reviewed diff in about five minutes.
Synara can feel unfamiliar on first launch because it is not an editor: it is a workspace that runs the AI coding agents you already pay for. This page is the shortest useful path through it. Every step links to a longer guide if you want the detail.
Install Synara and one provider
Download Synara from the install page. Then make sure at least one agent CLI is installed and signed in on your machine — for example Claude Code or Codex. Synara does not sell models; it discovers and drives the providers you already use.
Quick check: open a terminal and run the provider's CLI. If it works there, Synara can find it.
Add a project
Open Synara and add a local project folder. A Git repository is strongly recommended — it unlocks diffs, branches, worktrees, commits, and the pull-request workflow.
Start a task
Press mod+n to create a task in the active project. Pick a provider, a model, and an effort level in the composer. For your first task, work directly on the local project — worktrees matter once you run more than one task at a time.
Give the agent one concrete objective
State the outcome, the constraints, and how to verify it — not a vague theme:
Add an empty state to the pull-request list.
Reuse the existing shared panel components and run the focused browser test.Watch, steer, verify
While the agent works, the rest of the workspace stays one keystroke away: mod+j opens the terminal drawer for running tests yourself, the browser preview shows UI changes live, and the agent will pause when it needs an approval or an answer.
Review the diff, then commit
When the turn ends, read the diff before anything else. Commit only what you want to keep, push, and — if the repository is on GitHub — continue in the pull-request workspace.
That loop — task, objective, steer, review — is all of Synara. Everything else is a way to run more of those loops at once, safely.