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: Google Dialogflow providerId: google-dialogflow generated: '2026-09-12' method: searched source: https://cloud.google.com/dialogflow/quotas created: '2026-05-04' modified: '2026-09-12' description: >- Published Dialogflow quotas and limits, read from Google's own quotas page on 2026-09-12. This file replaces a 2026-05-04 scaffold whose numbers (10 rpm free / 100 rpm pro / 1000 rpm enterprise, X-RateLimit-* headers) were placeholder defaults and matched nothing Google publishes. Quotas apply per Google Cloud project and are shared across every application and IP using that project. Paid editions can request an increase; fixed limits cannot be raised. tags: - Conversational AI - Quotas - Rate Limiting headers: limit: null remaining: null reset: null retryAfter: null policy: null note: >- Dialogflow publishes NO rate-limit response headers. Remaining quota is not observable at runtime — an agent learns it is over only by receiving a 429, and must then follow the SLA's own Back-off Requirements (wait 1 second after the first error, doubling to a 32-second ceiling). Actual consumption is visible only in Cloud Monitoring (Conversational Agent LLM tokens quota usage and the per-quota metrics). responseCodes: throttled: 429 quotaExceeded: 429 canonicalStatus: RESOURCE_EXHAUSTED detailsType: google.rpc.QuotaFailure resetSemantics: perMinute: Refreshed every 60 seconds on the minute. perDay: Replenished daily at 12:00 AM Pacific Time. perMonth: Replenished on the 1st of each month at 12:00 AM Pacific Time. limit_count: 25 limits: - tier: Dialogflow CX Edition name: Text requests scope: project metric: requests_per_minute limit: 1200 timeFrame: minute applies: - Dialogflow CX API note: All DetectIntent, StreamingDetectIntent and FulfillIntent requests that do not contain audio. - tier: Dialogflow CX Edition name: Audio input/output requests scope: project metric: requests_per_minute limit: 600 timeFrame: minute applies: - Dialogflow CX API note: Speech-to-text, text-to-speech and telephony. - tier: Dialogflow CX Edition name: Audio query operations with Chirp 3 HD voice scope: project metric: requests_per_minute limit: 25 timeFrame: minute applies: - Dialogflow CX API note: On exceeding this, Chirp 3 HD voices may fall back to Neural or Standard voices rather than failing the request. - tier: Dialogflow CX Edition name: Generative features (ConversationalAgentLlmTokenConsumption) scope: project-and-region-and-model metric: tokens_per_minute limit: 600000 timeFrame: minute applies: - Dialogflow CX API note: >- Regional quota, tracked per combination of base model and region across Generators, Generative Fallback, Playbooks and Data stores. Covers gemini-1.5-pro, -pro-002, -flash, -flash-002, gemini-2.0-flash-001, -flash-lite-001, gemini-2.5-flash and -flash-lite. Multi-region increase requests must be filed against us-central1 (US/Global) or europe-west4 (EU). - tier: Dialogflow CX Edition name: Dialogflow CX Phone Gateway scope: project metric: phone_minutes_per_minute limit: 100 timeFrame: minute applies: - Dialogflow CX API - tier: Dialogflow CX Edition name: Design-time write requests scope: project metric: requests_per_minute limit: 60 timeFrame: minute applies: - Dialogflow CX API note: Calls that build or update an agent. This is the ceiling that constrains any agent doing bulk authoring. - tier: Dialogflow CX Edition name: Design-time read requests scope: project metric: requests_per_minute limit: 300 timeFrame: minute applies: - Dialogflow CX API - tier: Dialogflow CX Edition name: Other session requests scope: project metric: requests_per_minute limit: 100 timeFrame: minute applies: - Dialogflow CX API note: For example setting or getting session entities. - tier: Dialogflow ES Trial Edition name: Text requests scope: project metric: requests_per_minute limit: 180 timeFrame: minute applies: - Dialogflow ES API - tier: Dialogflow ES Trial Edition name: Audio input requests scope: project metric: requests_per_minute limit: 100 timeFrame: minute applies: - Dialogflow ES API note: Also capped at 1000 per day, 15,000 per month, and 60 seconds of audio per request. - tier: Dialogflow ES Trial Edition name: Audio input requests (daily) scope: project metric: requests_per_day limit: 1000 timeFrame: day applies: - Dialogflow ES API - tier: Dialogflow ES Trial Edition name: Audio input requests (monthly) scope: project metric: requests_per_month limit: 15000 timeFrame: month applies: - Dialogflow ES API - tier: Dialogflow ES Trial Edition name: Phone gateway scope: project metric: phone_minutes_per_minute limit: 3 timeFrame: minute applies: - Dialogflow ES API note: Also 30 phone minutes per day and 500 per month; number reserved for 30 days; no toll-free number. - tier: Dialogflow ES Trial Edition name: Mega agent requests scope: project metric: requests_per_day limit: 1000 timeFrame: day applies: - Dialogflow ES API - tier: Dialogflow ES Essentials Edition name: Text requests scope: project metric: requests_per_minute limit: 600 timeFrame: minute applies: - Dialogflow ES API - tier: Dialogflow ES Essentials Edition name: Audio input/output requests scope: project metric: requests_per_minute limit: 300 timeFrame: minute applies: - Dialogflow ES API note: Maximum 60 seconds of audio length per request. - tier: Dialogflow ES Essentials Edition name: Phone gateway scope: project metric: phone_minutes_per_minute limit: 100 timeFrame: minute applies: - Dialogflow ES API - tier: Dialogflow ES (both editions) name: Design-time write requests scope: project metric: requests_per_minute limit: 60 timeFrame: minute applies: - Dialogflow ES API - tier: Dialogflow ES (both editions) name: Design-time read requests scope: project metric: requests_per_minute limit: 60 timeFrame: minute applies: - Dialogflow ES API - tier: Dialogflow ES (both editions) name: Other session requests scope: project metric: requests_per_minute limit: 100 timeFrame: minute applies: - Dialogflow ES API note: For example setting or getting session entities, or updating/querying context. - tier: Agent Assist (Essentials Edition) name: Conversation and participant management operations scope: project metric: requests_per_minute limit: 300 timeFrame: minute applies: - Dialogflow ES API note: Not available on the Trial Edition. - tier: Agent Assist (Essentials Edition) name: AnalyzeContent / StreamingAnalyzeContent operations scope: project metric: requests_per_minute limit: 300 timeFrame: minute applies: - Dialogflow ES API - tier: Agent Assist (Essentials Edition) name: Concurrent bidi streaming sessions scope: project metric: concurrent_requests limit: 50 timeFrame: concurrent applies: - Dialogflow ES API - tier: Agent Assist (Essentials Edition) name: Conversation summary suggestion polling scope: project metric: requests_per_minute limit: 60 timeFrame: minute applies: - Dialogflow ES API - tier: Agent Assist (Essentials Edition) name: SearchKnowledge requests scope: project metric: requests_per_minute limit: 60 timeFrame: minute applies: - Dialogflow ES API fixedLimits: note: >- Fixed constraints on the service implementation. Unlike quotas these cannot be increased. Only the headline ones are reproduced here; the full table is on the quotas page. dialogflow_cx: - description: Maximum agents per Google Cloud project limit: 1000 - description: Maximum flows per agent limit: 50 - description: Maximum pages per flow limit: 250 - description: Maximum intents per agent limit: 10000 - description: Maximum entity types per agent limit: 250 - description: Maximum webhooks per agent limit: 100 - description: Maximum tools per agent limit: 100 - description: Maximum playbooks per agent limit: 50 - description: Maximum versions per playbook limit: 100 - description: Maximum test cases per agent limit: 2000 - description: Maximum environments per agent limit: 20 - description: Maximum changelog entries per agent limit: 20000 - description: Maximum training phrases per intent and per language limit: 2000 - description: Maximum entity entries per entity type limit: 30000 - description: Maximum steps per conversation limit: 100 - description: Maximum playbook invocations in one conversational turn limit: 3 - description: Maximum LLM calls per playbook invocation in the same turn limit: 10 dialogflow_es: - description: Maximum agents per region and per Google Cloud project limit: 1 - description: Maximum intents limit: 2000 - description: Maximum entity types limit: 250 - description: Maximum training phrases per intent and per language limit: 2000 - description: Maximum entity entries limit: 30000 - description: Maximum input contexts per intent limit: 5 - description: Maximum output contexts per intent limit: 30 - description: Maximum parameters per intent limit: 20 policies: - name: Back-off description: >- Stated in the SLA itself, not just in guidance: after the first error wait a minimum of 1 second, then increase the interval exponentially up to 32 seconds. Meeting this is a customer obligation for SLA credit eligibility. source: https://cloud.google.com/dialogflow/sla - name: Quota increase description: >- Paid editions can request an increase through the Cloud console quotas page. ES agents must be upgraded to the Essentials edition first. If the authenticating project differs from the project holding the agent, the increase must be requested on the consumer project. source: https://cloud.google.com/dialogflow/quotas - name: Session retention description: A session remains active and its data is stored for 30 minutes after the last request. source: https://cloud.google.com/dialogflow/quotas maintainers: - FN: Kin Lane email: kin@apievangelist.com