specification: API Commons Rate Limits specificationVersion: '0.1' schema: https://raw.githubusercontent.com/api-evangelist/interface-research/main/schema/api-commons.yml#/$defs/RateLimits provider: IBM Language Translator providerId: ibm-translate created: '2026-06-13' modified: '2026-06-13' reconciled: false tags: - Translation - IBM Watson - Rate Limiting - Deprecated description: >- IBM Watson Language Translator enforced per-instance rate limits. The Lite plan was limited to lower throughput compared to the Standard plan. IBM Cloud enforced HTTP 429 responses when limits were exceeded. The service was deprecated June 2023 and withdrawn December 2024. notes: >- IBM did not publish exhaustive numeric rate-limit tables in public documentation. Limits were enforced per service instance and varied by plan. Standard plan customers could contact IBM to discuss higher limits for production workloads. sources: - https://cloud.ibm.com/apidocs/language-translator - https://www.ibm.com/docs/en/watson responseCodes: throttled: 429 limits: - name: Lite Plan Concurrency scope: instance metric: requests-per-minute limit: 10 notes: Approximate; Lite plan had significantly lower request rate allowances. - name: Standard Plan Concurrency scope: instance metric: requests-per-minute limit: 50 notes: >- Approximate; Standard plan customers with high-volume needs could contact IBM for enterprise limits. - name: Character Limit Per Request scope: request metric: characters limit: 50000 notes: Single translate request capped at 50,000 characters. policies: - name: Backoff Strategy description: >- Use exponential backoff with jitter when receiving HTTP 429 responses. Honor any Retry-After header returned by the IBM Cloud API gateway. - name: Authentication description: >- All requests required IBM IAM authentication via a Bearer token obtained from https://iam.cloud.ibm.com/identity/token using an API key. maintainers: - FN: Kin Lane email: kin@apievangelist.com