generated: '2026-09-03' method: searched source: https://bitculator.com/en/documentation/api/v1 note: >- Limits published on the docs and pricing pages; header behavior confirmed by response examples throughout the API reference ("X-Quota-* on every call"). Two layers: a per-minute burst throttle per plan, then a monthly request quota. limit_count: 6 limits: - scope: per-account window: 1 month limit: 10000 plan: Free - scope: per-account window: 1 month limit: 50000 plan: Starter - scope: per-account window: 1 month limit: 250000 plan: Pro - scope: per-account window: 1 minute limit: 30 plan: Free kind: burst - scope: per-account window: 1 minute limit: 60 plan: Starter kind: burst - scope: per-account window: 1 minute limit: 120 plan: Pro kind: burst response_headers: - name: X-Quota-Limit detail: Monthly quota for the plan (e.g. 10000). - name: X-Quota-Used detail: Requests consumed this period (shown in every API reference response example). - name: X-Quota-Remaining detail: Requests left this period (shown in the quickstart example on the pricing page). - name: X-Quota-Reset detail: ISO-8601 instant the quota pool resets (e.g. 2026-07-01T00:00:00+00:00). exhaustion: error_code: rate_limited http_status: null detail: >- Over quota, the API returns the stable error envelope with code rate_limited "that tells the assistant when the pool resets" (MCP docs). The exact HTTP status on exhaustion is not explicitly stated in the public docs — recorded as unstated rather than guessed. Retry-After is not documented. introspection: operation: keyUsageQuota path: GET /api/v1/usage detail: >- Returns plan, limit, used, remaining, period_start/period_end, burst_per_minute, plus per-endpoint and per-token request tallies. quirk: >- The /usage response example in the docs shows "plan": "free" with "burst_per_minute": 60, while the pricing table and MCP docs both state the Free burst is 30/min — the example appears stale; the plan tables are taken as authoritative.