VICIdial for debt collection — managed hosting + TCPA scrubbing

Debt collection is the highest-volume, highest-regulated VICIdial use case in the US. Per-call TCPA scrubbing, FDCPA-aware dialer config, and 2–7 year recording retention are the table stakes. This page is the configuration the platform ships with — what you'd otherwise spend three weeks building on a self-hosted box.

Why managed VICIdial fits collections

Self-hosting works for shops with a full-time Linux + Asterisk admin. Most agencies grow past their initial setup faster than their ops capacity grows — managed hosting scales without that hire.

Concretely, the dialer needs to:

• Run predictive at 100–500 concurrent agents on bigger floors
• Scrub every dial against a TCPA litigator list before connecting
• Honor state-by-state call windows (NY, CA, MA, FL all have their own)
• Retain recordings for 2–7 years (state-specific)
• Audit-log every state transition for your compliance defense

The platform ships all of this. You bring the leads + your carrier.

Pre-dial TCPA scrubbing (per trunk)

The single biggest compliance risk in outbound collections is dialing a known litigator. The platform's TCPA feature checks every dial against Blacklist Alliance before Asterisk places the call — and lets you set the mode per outbound trunk.

Inherit   — use the server-wide BLA API key
Custom    — use this trunk's own key (different client, different contract)
Disabled  — skip BLA for this trunk

Per-trunk matters because a collections agency typically dials for multiple debt portfolios with different compliance contracts. Server-wide settings are too coarse.

You bring the BLA API key. We bring the AGI hook + 6-hour Redis cache + per-trunk override + the audit log row on every check.

See the TCPA Compliance feature page for the implementation detail.

Plan sizing for collections

| Floor size | Plan suggestion | | -------------- | --------------------------------- | | 10–25 agents | Growth (4 vCPU, 8GB) | | 25–75 agents | Pro (2 dedicated vCPU, 8GB) | | 75–200 agents | Business (4 dedicated vCPU, 16GB) | | 200–500 agents | Scale (8 dedicated vCPU, 32GB) | | 500+ agents | Enterprise sizing — talk to us |

Predictive at scale is RAM-hungry — vicidial_live_agents + vicidial_auto_calls hold a lot in memory. Budget ~200MB per concurrent agent on top of base.

Carrier choice

For collections, attestation matters more than per-minute cost. STIR/SHAKEN A-attestation is the difference between a 4% answer rate and a 12% answer rate.

The carriers we see collections shops run on:

• Bandwidth — direct US carrier, A-attestation, best wholesale rates at >500K min/mo
• Skyetel — wholesale-focused, US-only, very competitive at scale
• Telnyx — good for moderate volume + multi-region

Aggregator-routed carriers get worse answer rates over time as their reputation degrades. Pick direct-route, even if the per-minute is slightly higher.

VICIdial settings that matter

| Setting | Value | | ---------------------- | ------------------------------------------------ | | Local Call Time | 8am-9pm called party local | | Drop SLA | ≤ 2.0s (FCC 3% drop cap) | | Use Internal DNC | Y | | Recording Override | ALLFORCE (mandatory for litigation defense) | | Calls per Day per Lead | 1 (FDCPA harassment ceiling) | | Pre-dial AGI | tcpa-check.agi (platform-managed, do not edit) |

These land via the platform's default campaign profile when you create a collections campaign. You can override per-campaign in vicidial_admin.php; the platform never re-stomps your edits.

Recording retention

State + federal collection regs typically require 2–7 year retention. Plan storage:

• ~250KB per minute mono
• 100 agents × 4 hours/day × 250 days = ~360 GB / year

Two retention paths:

1. Platform Managed Backups — daily Hetzner snapshot + 7-day rolling retention. Fine for short-term rollback; not enough for litigation defense at year 3.
2. External Backups (BYO S3) — push recordings + DB to your own S3-compatible bucket (Backblaze, AWS S3, Wasabi, MinIO). Your bucket, your lifecycle policy. Hold for 2–7 years per your compliance schedule.

The two compose. Most serious collections shops run both.

Sub-user roles

• Owner — agency principal, full control
• Operator — floor supervisor, all operational access except billing
• Billing — back-office, sees invoices + payment methods, no agent management
• Viewer — read-only auditors, compliance staff

Every action lands in the audit log with actor + timestamp. When the suit lands, your discovery is already prepared.

Recorded SSH sessions

Every SSH session into your box is captured as a replayable asciinema cast. For SOC2 + customer-attestation reviews that ask "who touched the production system, and what did they do" — replay the actual session. This is a normal feature for VICIfast; on self-hosted you'd assemble it from auditd + tlog + journal-based heuristics.

Why collections teams switch from self-hosted

• Server up in under 60 seconds vs hours-to-days when you compile Asterisk yourself
• A real dashboard instead of "email a tech / SSH in"
• Audit log ready before the FDCPA suit lands, not after
• Per-trunk TCPA scrubbing ships out of the box
• Transparent pricing — per server, no per-agent licensing
• Migration is free — we move your existing vicidial_* schema + recordings; see /migrate-from-vicibox

Get started

Start a free trial — card on file, no charge for 7 days. For floor sizes above 100 agents, talk to us first so we can size the right plan + region.