{
  "schemaVersion": "2026-05-14.tool-review.v1",
  "slug": "deepgram",
  "name": "Deepgram",
  "category": "voice-speech",
  "verdict": {
    "label": "Recommended",
    "tone": "recommended",
    "summary": "Use Deepgram when low-latency speech-to-text is central to the product experience."
  },
  "scores": {
    "dx": 84,
    "ax": 76,
    "production": 82,
    "pricing": 71,
    "performance": 90
  },
  "pricingTier": "usage-based",
  "agentReadiness": "partial",
  "timeToFirstSuccessMinutes": 30,
  "recommendedFor": [
    "voice-agent",
    "prototype-to-demo"
  ],
  "avoidWhen": [
    "text-only agents",
    "offline-only transcription needs",
    "teams that cannot model audio usage costs"
  ],
  "evidence": {
    "built": "Tested voice/speech API integration paths for AI agent demos and content workflows.",
    "testedScenario": "Realtime transcription setup with latency and developer onboarding checks.",
    "methodology": [
      "Checked quickstart clarity",
      "Reviewed SDK fit",
      "Mapped realtime architecture",
      "Compared latency-sensitive caveats"
    ]
  },
  "evidenceProfile": {
    "level": "strong",
    "artifacts": [
      {
        "kind": "human-review",
        "label": "Human review page",
        "href": "/blueprints/reviews/deepgram",
        "public": true
      },
      {
        "kind": "agent-json",
        "label": "Agent JSON verdict",
        "href": "/blueprints/reviews/deepgram.json",
        "public": true
      },
      {
        "kind": "compare-view",
        "label": "Compare with alternatives",
        "href": "/blueprints/reviews/compare?tools=deepgram,cursor,claude-code",
        "public": true
      }
    ],
    "limitations": [
      "Scores reflect Neurl hands-on evidence and should be re-verified before procurement or high-risk production adoption.",
      "Pricing, limits, model defaults, and product policies can change quickly; use freshness dates and vendor docs before final rollout."
    ],
    "confidenceSignals": [
      "Tested scenario: Realtime transcription setup with latency and developer onboarding checks.",
      "4 methodology checks",
      "Last verified: 2026-05-14",
      "2 agent safe-use notes"
    ],
    "agentEvidenceSummary": "Deepgram was tested in scenario \"Realtime transcription setup with latency and developer onboarding checks.\" and last verified on 2026-05-14. Use the human review, agent JSON verdict, and compare view before trusting the recommendation."
  },
  "freshness": {
    "lastTestedAt": "2026-04-18",
    "lastVerifiedAt": "2026-05-14",
    "staleAfterDays": 90,
    "scoreDiffLog": [
      "2026-05-14: added to voice-agent task filter"
    ],
    "changelogPulse": "Speech APIs are improving quickly; re-test realtime behavior before major launches."
  },
  "agent": {
    "skillText": "Use Deepgram when the task needs speech-to-text, realtime transcription, or a voice-agent input pipeline. Avoid it when the product is text-only or audio usage costs cannot be estimated.",
    "manifestSnippet": {
      "name": "deepgram",
      "useWhen": [
        "voice-agent",
        "realtime transcription",
        "speech-to-text"
      ],
      "avoidWhen": [
        "text-only agent",
        "offline-only transcription",
        "unknown audio budget"
      ],
      "requiredContext": [
        "audio source",
        "latency target",
        "language requirements",
        "expected minutes"
      ],
      "confidence": "medium"
    },
    "safeUseNotes": [
      "Prototype with real audio samples",
      "Design fallback states for transcript uncertainty"
    ]
  }
}