The TypeScript package ships the tinyhumans command. It mirrors the typed SDK:
every namespace method is a tinyhumans <namespace> <command> subcommand
(156 commands across 17 namespaces), with a raw escape hatch for anything not
yet typed.
tinyhumans # top-level usage + namespace list
tinyhumans list # every command, grouped by namespace
tinyhumans auth # list the auth namespace's commands
tinyhumans auth verify-email --help # show a command's signatureEach command maps to its typed method's parameters:
- positional params are passed in order (or as
--<name> VALUE), - the request body comes from
--body JSONor piped stdin, - query params come from repeatable
--query key=value.
# GET /auth/me
tinyhumans auth me --token "$TINYHUMANS_TOKEN"
# GET /auth/email/verify?token=... (token is a positional param)
tinyhumans auth verify-email tok_123
# PUT /teams/:teamId with a JSON body
tinyhumans teams update-team team_9 --body '{"name":"New name"}' --token "$TINYHUMANS_TOKEN"
# POST /openai/v1/chat/completions (body via stdin; returns raw provider JSON)
echo '{"model":"gpt-4.1-mini","messages":[]}' \
| tinyhumans inference create-chat-completion --token "$TINYHUMANS_TOKEN"
# Query params
tinyhumans inference list-models --query limit=5tinyhumans swagger --base-url https://api.tinyhumans.ai
tinyhumans raw get /auth/me --token "$TINYHUMANS_TOKEN"
echo '{"email":"you@example.com"}' | tinyhumans raw post /auth/email/send-link--base-url URL--token TOKEN--api-key KEY--admin-service-token TOKEN--raw— return the full{ success, data }envelope instead of unwrapping--json— compact JSON output (default is pretty-printed)
TINYHUMANS_BASE_URLTINYHUMANS_TOKENTINYHUMANS_API_KEYTINYHUMANS_ADMIN_SERVICE_TOKEN
The subcommand table is generated from the typed API modules into
src/cli-commands.ts (run automatically by build):
pnpm --filter @tinyhumansai/sdk gen:cli