Quick Start

Use this page for the shortest credible path. If you want the guided version, read [[Getting Started|Getting-Started]].

1. Verify The Install

boundline doctor --install

Do not skip this. It tells you whether the installed Boundline binary and the documented Canon pairing are ready.

2. Initialize The Workspace

cd <workspace>
boundline init --assistant codex
boundline config set-semantic-acceleration --scope workspace --policy local
boundline index status --workspace .

That bootstraps .boundline/ and, when requested, the repo-local assistant surface for the selected host.

3. Optional Provider Auth

Use this only when the chosen runtime needs a stored provider credential:

boundline models auth login --provider github-copilot
boundline models auth status

These credentials are user-scoped, not repository-scoped.

4. Optional Readiness Probe

Use probe when you want a read-only answer before starting or resuming:

boundline probe

If probe says bootstrap is still required, go back to init. If it says repair is needed, follow the printed action. If it says the session is ready, continue with the normal loop.

If the workspace uses local semantic retrieval, add:

boundline index refresh --workspace .
boundline index doctor --workspace .

Use refresh to rebuild bounded local evidence and doctor when the manifest, tracked-file hygiene, or vector capability looks wrong.

5. Run One Bounded Session

boundline goal --goal "Fix the failing add test"
boundline plan
boundline run
boundline status
boundline inspect

This is the primary product path.

If you want the shortest path after init, you can use:

boundline run --goal "Fix the failing add test"

Treat that as a fast path, not the default mental model.

For an explicit planning seed during bootstrap, use planning=copilot:gpt-4o:

boundline init --assistant copilot --route planning=copilot:gpt-4o

What This Path Gives You

• explicit session state under .boundline/
• a bounded plan built from repository evidence
• read-side status and trace inspection
• explicit stop conditions when context, validation, or governance is not ready

Next step: read [[Daily Operating Guide|Daily-Operating-Guide]] when you want the normal loop, follow-through, and recovery behavior in more detail.