spec: API Commons Rate Limits 0.1
provider: Gladia
providerId: gladia
created: 2026-06-12
modified: 2026-06-12
url: https://docs.gladia.io/chapters/limits-and-specifications/concurrency
throttling:
  httpStatusCode: 429
  retryAfterHeader: Retry-After
  description: >
    Gladia enforces API-level rate limits applied uniformly across all users
    to prevent performance degradation. Requests that exceed rate limits
    receive an HTTP 429 response. Exact per-minute call thresholds are
    not publicly documented; contact Gladia support for specifics.
limits:
  - name: Free tier monthly usage cap
    scope: account
    plan: Starter (free quota)
    type: quota
    unit: hours
    limit: 10
    period: monthly
    description: >
      Free-tier users are limited to 10 hours of transcription per month
      across async and real-time combined.

  - name: Free tier async concurrency
    scope: account
    plan: Starter (free quota)
    type: concurrency
    unit: simultaneous_requests
    limit: 3
    description: Maximum simultaneous pre-recorded (async) transcriptions for free-tier users.

  - name: Free tier real-time concurrency
    scope: account
    plan: Starter (free quota)
    type: concurrency
    unit: simultaneous_sessions
    limit: 1
    description: Maximum simultaneous live transcription sessions for free-tier users.

  - name: Paid async concurrency
    scope: account
    plan: Starter (paid) / Growth
    type: concurrency
    unit: simultaneous_requests
    limit: 25
    description: Maximum simultaneous pre-recorded (async) transcriptions for paid-tier users.

  - name: Paid real-time concurrency
    scope: account
    plan: Starter (paid) / Growth
    type: concurrency
    unit: simultaneous_sessions
    limit: 30
    description: Maximum simultaneous live transcription sessions for paid-tier users.

  - name: Paid async request queue depth
    scope: account
    plan: Starter (paid) / Growth
    type: queue
    unit: requests
    limit: 300
    description: >
      Paid users can queue up to 300 async transcription requests; only 25
      will be actively processed concurrently at any given time.

  - name: Real-time session duration
    scope: session
    plan: All
    type: duration
    unit: hours
    limit: 3
    description: >
      A single live (real-time) WebSocket transcription session cannot
      exceed 3 hours in duration regardless of plan.

  - name: Enterprise concurrency
    scope: account
    plan: Enterprise
    type: concurrency
    unit: simultaneous_requests
    limit: unlimited
    description: Enterprise accounts receive on-demand concurrency negotiated with Gladia sales.