@mirage-cli/radar-cli

ViewEngine Radar CLI — org-scoped access to the ViewEngine AI Visibility API at radar.viewengine.ai. Create and onboard projects; read queries, query-results, execution-jobs, and credits; manage game plans; switch between organizations.

bun add -g @mirage-cli/radar-cli
radar login                    # Clerk OAuth (browser)
radar login --api-key sk_...   # headless / CI

radar orgs list                # orgs you belong to
radar orgs use <id-or-slug>    # set the active org (persisted)

radar projects list --search acme
radar projects create example.com
radar onboarding analyze <projectId> --domain example.com --verbose
radar onboarding save <projectId> --section business --data @business.json
radar onboarding generate-queries <projectId> --profile @profile.json
radar onboarding complete <projectId> --queries @queries.json
radar queries list --project-id <id>
radar game-plans list --status open
radar game-plans complete-action <planId> <actionIndex>
radar results list --provider perplexity --query-id <id>
radar jobs list --status running
radar credits list
radar export game-plans > game-plans.json

Commands

• projects {create, list, get} — creation uses the onboarding-aware route; renaming and descriptions remain a separate concern until that dashboard update contract is represented in this repo
• onboarding {create, status, analyze, save, generate-queries, complete} — headless project setup; JSON options accept inline JSON or @file, and the two SSE commands buffer output until completion (--timeout, --verbose)
• queries {list}
• game-plans {list, get, update, complete-action}
• results {list, get} (query-results)
• jobs {list} (execution-jobs)
• credits {list}
• export <entity> — page through every row of an entity
• metrics {overview, project, brands, sources, trends, heatmap} — server-side aggregates (dashboard-identical numbers); --days/--from/--until/--platforms windowing, --format table|csv|json, overview --compare for period-over-period deltas
• export-results — full-history NDJSON streamed server-side (--since resume, -o <file>)
• orgs {list, use, current, clear} — multi-tenant switching
• login / whoami / logout

All list commands accept --page, --limit, --sort <field>, --dir <asc|desc>. Output is JSON (ideal for piping / bot consumption).

Auth & org scoping

• radar login — Clerk OAuth (PKCE + RFC 7591 DCR + loopback). Persists to ~/.config/radar/session.json (0600).
• radar login --api-key sk_... — Clerk machine API key for headless use.
• Env override: RADAR_API_KEY (preferred) or RADAR_OAUTH_ACCESS_TOKEN. RADAR_API_BASE_URL overrides the base URL.
• Active org: radar orgs use <id-or-slug> persists it; override per-command with --org <id-or-slug> or the RADAR_ACTIVE_ORG_ID env var. It rides on every request as the X-Active-Org-Id header, which the API resolves by id or slug server-side.
• Shared projects: projects, queries, results, and game-plans scoped to the active org include projects shared into it (agency access), not just ones it owns. Each project row carries an access field — "owner" (your org owns/pays) or "shared" (shared in for view/run/edit; billing hidden). Jobs and credits stay owner-org-only.

Programmatic use

import { buildProgram, ApiClient, loadSession } from "@mirage-cli/radar-cli";

// As a Commander program (drives the same CLI surface):
const program = buildProgram();
await program.parseAsync(["node", "radar", "projects", "list"]);

// As a typed API client:
const session = loadSession();
if (session) {
  const client = new ApiClient(session);
  const { rows, total } = await client.list("projects", { search: "acme" });
  const plan = await client.get("game-plans", "<id>");
}

Drop-in for mirage: see @mirage-cli/radar.