generated: '2026-07-21'
method: searched
source: openapi/textcortex-openapi.json + docs.textcortex.com
summary: >-
  Cross-cutting request/response semantics for the OpenAI-compatible TextCortex API
  (https://api.textcortex.com/v1). The surface mirrors the OpenAI Chat Completions and
  Responses contracts, so OpenAI conventions apply.
authentication:
  style: http-bearer
  header: 'Authorization: Bearer <api_key>'
  ref: authentication/textcortex-authentication.yml
error_envelope:
  shape: openai-error
  format: application/json
  example_field_path: error.message
  fields: [error.message, error.type, error.code, error.param]
  types: [invalid_request_error, server_error]
  note: OpenAI-style error object, not RFC 9457 problem+json.
  ref: errors/textcortex-problem-types.yml
streaming:
  supported: true
  mechanism: server-sent-events
  content_type: text/event-stream
  toggled_by: 'request field: stream=true'
  operations: [createChatCompletion, createResponse]
idempotency:
  supported: false
  note: No idempotency-key header or parameter is documented in the OpenAPI or docs.
pagination:
  supported: false
  note: >-
    GET /models returns an OpenAI-style {object:"list", data:[...]} envelope with no
    cursor/offset paging parameters. No other collection endpoints.
versioning:
  scheme: uri-path + calendar-versioned info.version
  current_path: /v1
  spec_version: 1.2026.07.20
  ref: lifecycle/textcortex-lifecycle.yml
rate_limiting:
  signal: undocumented
  note: >-
    No rate-limit response headers are declared in the OpenAPI. Usage is metered by API
    credits; 402 "Insufficient API credits" is returned when the account balance is
    exhausted (see GET /balance).
data_residency:
  note: >-
    Models expose served_from_country_code; TextCortex advertises EU-hosted, GDPR-compliant
    infrastructure with per-model Europe overrides.