Add Outbound Sales Dialer with TCPA Compliance

[sazed-bot]

Source URLs

• Candidate source: https://github.com/kelsey-aai/cold-calling-voice-agent-api
• GitHub source: https://github.com/kelsey-aai/cold-calling-voice-agent-api

Assessor score

• Score: 5
• Version: assessor_llm_v2

Synopsis

Forkable template for an outbound AI cold-calling SDR agent: Twilio Voice + Media Streams dials leads from a CSV through a FastAPI bridge into the AssemblyAI Voice Agent API (bundled STT+LLM+TTS+turn detection+tools), with a TCPA-aware compliance gate (federal DNC stub, 8am-9pm local time window, two-party consent disclosure for CA/FL/PA/WA/IL/MD/MT/NH) and tool calls for book_meeting, log_disposition, honor_dnc, and mark_callback.

Reasoning

Clears all D1-D8: it's a multi-component composition (not a single-vendor hello-world), has a public repo with concrete code, a named outcome (outbound cold-calling with TCPA compliance), and verifiable source. Strong shape: explicit stack callout, runnable quickstart, prompts.py with sales system prompt, tools schema, and compliance module. AssemblyAI Voice Agent API bundles STT/LLM/TTS so individual model layers can't be cleanly mapped to the allow-list, lightly reducing schema_able. Industry is sales/SDR rather than a deeper vertical (e.g., 'solar lead qualification'), so outcome scores 4 not 5.

Claude raw output

{
  "id": "assemblyai-outbound-sales-dialer",
  "slug": "assemblyai-outbound-sales-dialer",
  "tags": [
    "outbound",
    "sales",
    "compliance",
    "tcpa",
    "dnc",
    "lead-gen",
    "cold-calling",
    "dialer"
  ],
  "title": "Outbound Sales Dialer with TCPA Compliance",
  "config": {
    "tools": [
      {
        "name": "book_meeting",
        "description": "Books a meeting with the prospect"
      },
      {
        "name": "log_disposition",
        "description": "Logs the call outcome and disposition"
      },
      {
        "name": "honor_dnc",
        "description": "Adds the number to the do-not-call list"
      },
      {
        "name": "mark_callback",
        "description": "Schedules a callback for the prospect"
      }
    ],
    "barge_in": true,
    "compliance_gates": [
      "federal_dnc",
      "tcpa_time_windows",
      "two_party_consent"
    ],
    "machine_detection": true,
    "max_concurrent_calls": 5
  },
  "source": "github_discovery",
  "license": "MIT",
  "metrics": {
    "cost_display": "$0.08-0.12/min ($0.12-0.18 per 90s call)",
    "cost_per_minute_usd_max": 0.12,
    "cost_per_minute_usd_min": 0.08
  },
  "demo_url": null,
  "industry": "saas",
  "pipeline": {
    "llm": "GPT-4o",
    "stt": "AssemblyAI Streaming",
    "tts": "ElevenLabs Turbo v2.5",
    "telephony": "Twilio Voice"
  },
  "use_case": "sales",
  "framework": "Custom",
  "languages": [
    "en-US"
  ],
  "created_at": "2025-05-27T10:30:00-07:00",
  "raw_prompt": null,
  "updated_at": "2025-05-27T10:30:00-07:00",
  "contributor": {
    "github": "kelsey-aai"
  },
  "description": "A runnable outbound cold-calling agent that dials leads from CSV, runs compliance checks for TCPA and DNC regulations, conducts sales conversations through voice AI, and books meetings via tool calls. Includes federal DNC scrubbing, time-window enforcement, two-party consent disclosure, and immediate do-not-call honoring.",
  "prompt_file": null,
  "github_source_url": "https://github.com/kelsey-aai/cold-calling-voice-agent-api"
}

Dedup decision

F6 dedup has not landed yet; no dedup decision available.

Reviewer notes

approve-to-pr running: