Connect with files: S3 and SFTP
Level 0 is a real integration, not a consolation prize. If your world is spreadsheets, your spreadsheets are the interface.
The drop
s3://switchboard-ingest/yourco/2026-W40/ catalog.csv # package_id,type,network,selling_title,days,start,end,currencies,tiers rates.csv # package_id,week,class,advertiser_class,spot_length,tier,clearance,gross_rate_unit avails.csv # package_id,week,date,state,spots_available,spots_total,sellout_level,as_of # leave date blank for week-grain rows; fill it to publish a slot-grain row manifest.json # schema_version, sha256 per file, effective window
Register each drop with POST /v1/me/files/manifest. Accepted formats: csv, parquet, xlsx, and json.
- SFTP works identically to S3. Idempotency is by file hash: re-dropping the same manifest is a no-op.
- Validation and the integrity rule run on ingest; an
errors.jsonreport lands next to your drop within minutes, or your catalog goes live. - Documents you receive back (order exports, confirmations) are hash-bound renderings of the authoritative JSON: the sha256 in the manifest is the tamper-check.
- Orders flow to your approval queue, and decisions can return by console click, email-gateway reply, or webhook, whichever your team actually uses.
A national seller can be live on files this week and move rates or avails to an API next quarter without touching anything else. Direction is per capability.