generated: '2026-07-21'
method: searched
source: https://thetokencompany.com/docs ; github.com/TheTokenCompany/the-token-company-python
authentication:
  style: bearer
  header: Authorization
  ref: authentication/the-token-company-authentication.yml
idempotency:
  supported: false
  note: Compression endpoints are effectively idempotent for identical input (no side effects), but the API documents no Idempotency-Key contract.
pagination:
  supported: false
  note: No list endpoints; responses are single objects. Web search caps results via max_results.
request_payload:
  content_type: application/json
  gzip: >-
    Optional gzip of the request body (Content-Encoding: gzip) for large inputs;
    the SDK exposes gzip=True (up to ~2.2x faster on 1M+ token inputs).
content_protection:
  mechanism: <ttc_safe>...</ttc_safe> tags
  note: Text wrapped in ttc_safe tags is excluded from compression and passes through unchanged.
usage_tagging:
  field: app_id
  note: Optional application identifier (max 255 chars) tags requests for usage tracking, per-request or client-level.
compression_control:
  field: aggressiveness
  range: 0.0 - 1.0
  default: 0.2
  note: Chat compression also accepts a per-role aggressiveness map.
versioning:
  ref: lifecycle/the-token-company-lifecycle.yml
  note: API is /v1; compression capability is versioned by model (bear-2 latest).
error_envelope:
  ref: errors/the-token-company-problem-types.yml
  note: HTTP status codes mapped to typed SDK exceptions.
rate_limiting:
  signal: HTTP 429 (RateLimitError)
  note: No documented rate-limit response headers; retry with backoff.
data_retention:
  policy: zero data retention
  docs: https://thetokencompany.com/docs/data-retention