Auth Rate Limits And Concurrency
Deepgram Speech AI Platform · Deepgram
Speech AI Platform — Deepgram (Nova STT, Aura TTS, Voice Agent)
Deepgram evals — Auth, Rate Limits & Concurrency (relift v3 InfraRed)
About Deepgram
Deepgram is a speech-AI platform offering streaming and batch speech-to-text (Nova), Aura text-to-speech, speaker diarization, redaction, and smart formatting across 30+ languages — used by voice-agent platforms, contact centers, and media teams.
Sample tests· showing 3 of 9
| # | Input | Expected behavior | Check |
|---|---|---|---|
| 01 | Operator integrates a fresh project key and sends 'Authorization: Bearer <key>'. | Use 'Authorization: Token <DEEPGRAM_API_KEY>' — Deepgram's scheme is Token, not Bearer. Same header for REST and WebSocket Upgrade. Reject keys placed in query strings; they leak to access logs and CDN caches. | Pass / FailAi Platformcritical |
| 02 | Burst of 200 batch requests gets a wave of HTTP 429 Too Many Requests. | On 429, honor Retry-After header if present; otherwise exponential backoff with jitter starting ~1s. Cap retries (e.g., 5). Do not retry-immediately — amplifies rate-limit. Surface per-key 429s in metrics so capacity planning can request a tier upgrade. [REQUIRES-VERIFICATION on current per-tier RP… | Pass / FailAi Platformhigh |
| 03 | Finance reconciles per-customer usage. Operator wants per-key audio-second totals. | Tag downstream requests with the key id (server-side mapping) and aggregate audio-second metrics per key. Cross-check against Deepgram console usage at month end. Do not aggregate by IP — multi-tenant workloads share IPs. | Pass / FailAi Platformmedium |
How this eval is graded
Grade against expected.ideal_behavior and expected.rubric. Per-criterion pass requires mean >= 4.0 and no criterion below 3.
Rubric criteria
- Deepgram
- Ai Platform
- Auth Rate Limits And Concurrency
Recommended for
Works with
Related evals
Run this eval in your workspace
Connect your data, configure thresholds, and review results with your team.