> ## Documentation Index > Fetch the complete documentation index at: https://atlas.rebuildingus.org/docs/llms.txt > Use this file to discover all available pages before exploring further. # Scout CLI > Run Atlas discovery locally, review source evidence, and sync selected results when they are ready. Scout is the local Atlas discovery CLI. It can run on its own against public URLs, files of URLs, stored article corpora, or place and issue queries. When connected to Atlas, it can sync reviewed local research into a public review flow or a private workspace. Use Scout when you need local control over discovery: a specific city, a narrow research question, a repeatable script, or a reviewable source trail before anything enters Atlas. ## What Scout Does Locally * Fetches public sources and extracts possible civic actors. * Reads structured public resources such as CSV, TSV, pipe-delimited text, and ZIP files that contain those formats. * Extracts entries from stored news article corpora during normal `scout run --location ... --issues ...` runs. * Follows same-domain links when configured. * Uses a local or configured model provider to structure results. * Stores local runs, pages, and entries in a local Scout database. * Lets you inspect the evidence before syncing. * Exports source-linked entry files for review, audits, and handoff. * Stores dated article corpora separately from extracted entries. Scout is not only an Atlas API client. You can run local discovery without syncing anything. Atlas becomes useful when you want shared review, workspace visibility, or contribution receipts. ## Ways To Use Scout | Workflow | Best for | Atlas account | | ------------------------- | ----------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------- | | Known URLs | Inspecting known public pages and saving local results. | Not required | | Stored article corpus | Turning dated article records into sourced people, organizations, initiatives, campaigns, and events. | Not required | | Place and issue discovery | Finding sources for a place and issue set. | Not required, but search connection is needed for new-source discovery | | Sync to Atlas | Moving reviewed local results into Atlas review or a workspace. | Required | | Worker mode | Letting Atlas assign discovery jobs to this computer. | Required | ## Trust Boundary Scout produces research results, not verified public claims. A completed run means Scout finished processing sources. It does not mean every extracted person, organization, initiative, campaign, or event is ready for outreach, publication, funding review, or import. Keep the source URL, title, date, source type, and extraction context with any claim you use. The source evidence is what makes Scout results useful. Currentness is a separate question from source-backing. A historical public record can prove that something was true or reported at the time without proving that it is true now. Treat old filings, archives, rosters, and transaction files as historical evidence unless their source date and context support a current claim. Use `scout entries stats` to check source-backed totals, exact unique people, duplicate surplus, and source coverage before syncing or exporting. Use entry samples and source context to decide whether the run is current enough for the use case. ## Start Here Install the CLI and confirm the command is available. Configure Atlas login, local models, search keys, and readiness checks. Run URLs, files, article corpus recovery, and place-and-issue discovery. Collect dated news article records that normal Scout runs can extract from. Inspect local results before exporting or syncing. Sync reviewed runs to public review or a workspace.