Insert or upsert rows into a Supabase table via the Data REST API (PostgREST) at /rest/v1/{table}, controlling whether the response body is returned and how conflicts are resolved.

Documented steps

1. Get your Project URL and a key from Settings > API Keys. Use the publishable key (sb_publishable_...) for client-side inserts guarded by RLS, or the secret key (sb_secret_...) only from a trusted server for elevated access that bypasses RLS. Legacy anon/service_role JWT keys still work identically until you disable them.
2. Single insert: POST https://<project_ref>.supabase.co/rest/v1/{table} with headers 'apikey: <KEY>', 'Content-Type: application/json', and body as a plain JSON object, e.g. curl "https://<ref>.supabase.co/rest/v1/todos" -X POST -H "apikey: <KEY>" -H "Content-Type: application/json" -d '{"task":"write docs"}'. If acting as a logged-in user, also send 'Authorization: Bearer <user_access_token>' (the user's Supabase Auth session JWT, not the project key) so auth.uid() resolves for RLS.
3. By default writes return no body (Prefer: return=minimal is the default for write requests). Add header 'Prefer: return=representation' to get the inserted row(s) back, or 'Prefer: return=headers-only' for a Location header like 'Location: /todos?id=eq.34'.
4. Bulk insert: POST an array of JSON objects with uniform keys, e.g. -d '[{"task":"a"},{"task":"b"}]', which runs as a single INSERT statement server-side.
5. Upsert by primary key: POST with header 'Prefer: resolution=merge-duplicates' (or 'resolution=ignore-duplicates' to silently skip conflicts). Upsert matches on the primary key by default, so include all primary key columns in every row payload.
6. Upsert on a different unique column: add the on_conflict query parameter, e.g. POST /rest/v1/employees?on_conflict=name with 'Prefer: resolution=merge-duplicates'.
7. For surrogate keys with some rows missing an id (new inserts mixed with existing-id updates), combine 'Prefer: resolution=merge-duplicates, missing=default' with ?columns=id,name,salary so omitted non-key columns fall back to their DEFAULT instead of being set to null.
8. Docs: https://supabase.com/docs/guides/api/creating-routes https://supabase.com/docs/guides/api/api-keys https://docs.postgrest.org/en/latest/references/api/tables_views.html https://docs.postgrest.org/en/latest/references/api/preferences.html