Pricing and credits

Talqing charges one platform fee, debited from a prepaid USD credit balance per workspace. Provider spend (OpenAI, Deepgram, ElevenLabs, Anam…) is billed to your own accounts on your own provider keys; Talqing reports it per call and never collects it.

The platform fee

Channel	Fee	Charged on
`voice`	$0.0035 per minute	Phone, browser, WhatsApp and media-stream calls
`video`	$0.01 per minute	Browser calls with an avatar
`text`	$0.0001 per answered message	Chats over the API, Telegram and WhatsApp

Minutes are the call’s duration, billed to the second. No per-call minimum, connection fee or subscription.
An answered message is one the agent finished replying to. A message superseded before it was answered, or whose turn failed, is free. Time a chat sits open is not billed.

The fee is waived on failed calls

A call ending with one of these close reasons is not charged the platform fee (your provider spend on it still stands): error, stale, job_shutdown, job crashed, orphaned, unknown, avatar_start_failed, invalid_dispatch_metadata, missing_dispatch_metadata, unknown_tenant, missing_published_definition, rejected, insufficient_credits, and any reason ending in _failed.

Every other call is charged, including a caller who hangs up after two seconds, a call answered by voicemail, and a call with no close reason at all. See close reasons.

Reading the balance

Any member can read it.

curl https://api.talqing.com/v1/billing/credits \
  -H "Authorization: Bearer $TALQING_API_KEY"

const credits = await talqing.billing.credits.get();
console.log(credits.balance, credits.voice_minutes_remaining);

credits = talqing.billing.credits.get()
print(credits["balance"], credits["voice_minutes_remaining"])

{
  "balance": 42.5,
  "currency": "USD",
  "voice_minutes_remaining": 12142,
  "video_minutes_remaining": 4250,
  "low_balance_threshold": 2.0,
  "packs": [
    { "id": "usd_20", "amount": 20.0, "voice_minutes": 5714 },
    { "id": "usd_50", "amount": 50.0, "voice_minutes": 14285 }
  ],
  "topup_status": null
}

Field	Meaning
`balance`	USD. Can be slightly negative (see running out)
`voice_minutes_remaining`, `video_minutes_remaining`	Balance divided by that channel’s fee. Platform fee only; provider spend is extra
`low_balance_threshold`	Below this ($2) the dashboard shows a warning. It never blocks
`packs`	The packs on sale, with how many voice minutes each buys
`topup_status`	Your latest purchase: `pending`, `paid`, `failed`, `cancelled`, `refunded`, or `null`

The workspace created at your first sign-in gets $10 once. A workspace you create later starts at $0.

Buying credit

ADMIN only. Credit comes in fixed packs: usd_20, usd_50, usd_100, usd_200, usd_500, usd_1000. The id is the amount credited, regardless of currency conversion, tax or processor fees. In the dashboard: Organization → Settings → Billing.

curl -X POST https://api.talqing.com/v1/billing/credits/checkout \
  -H "Authorization: Bearer $TALQING_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{ "pack": "usd_50" }'

const checkout = await talqing.billing.credits.checkout({
  creditCheckoutRequest: { pack: "usd_50" },
});
console.log(checkout.checkout_url, checkout.topup_id);

checkout = talqing.billing.credits.checkout(pack="usd_50")
print(checkout["checkout_url"], checkout["topup_id"])

Open checkout_url in a browser; card details never touch Talqing. After paying, read the balance again: that read applies a confirmed payment, and topup_status shows whether the processor has confirmed it yet.

Running out

The only check is balance > 0. There is no reservation and no spend cap.

Where	At zero balance
`POST /v1/calls/token`, `POST /v1/calls/outbound`, `POST /v1/chats`, resuming a call	HTTP 402
Inbound phone call, media stream or WhatsApp call	Refused, recorded with `close_reason: "insufficient_credits"` (never charged)
A Telegram or WhatsApp message that would start a chat, or a message to an idle chat	The agent does not run; over the API the turn returns `status: "error"`
A running call batch	Dials nothing and shows why; resumes by itself once credit is added. No retry attempt is spent

The 402 uses the standard error envelope:

{
  "detail": {
    "message": "This workspace is out of credits. Add credits in Settings → Billing to start new calls. Calls already in progress are unaffected.",
    "errors": []
  }
}

A call already running is never cut off, and an active chat is not re-checked on every message. The balance can therefore go slightly negative; the next top-up pays that off first.

The ledger

GET /v1/billing/credits/ledger lists every balance movement, newest first, each with amount, balance_after, and session_id or topup_id. A usage row also carries its session’s channel (voice, video or text) and conversation_id.

`kind`	What it is
`signup_grant`	The one-time signup credit
`purchase`	A pack you bought (`topup_id`)
`usage`	A call’s fee when it ends, or a chat’s fee as messages are answered (`session_id`). A long chat has several rows summing to its fee. Failed calls write no row
`refund`	A payment reversed by the processor (`topup_id`)
`adjustment`	A manual correction, explained in `note`

curl "https://api.talqing.com/v1/billing/credits/ledger?limit=50" \
  -H "Authorization: Bearer $TALQING_API_KEY"

Estimating cost per minute

Provider spend is usually far larger than the platform fee, so estimate it before running at volume.

Dashboard: the agent editor’s right rail shows a live per-minute estimate for the draft, split into STT, LLM, TTS (or Realtime), Avatar and Platform.
API: GET /v1/catalog returns each model’s provider rates plus the assumed usage behind an estimate: cost_estimate_usage_per_minute (per spoken minute: 800 LLM input and 250 output tokens, 900 TTS characters, 60 s of streaming or 20 s of batch STT) and cost_estimate_usage_per_text_message (800 input and 250 output tokens per reply). Multiply the rates by those quantities and add platform_fee_per_minute or platform_fee_per_message.

These are estimates. The bill comes from what each provider metered on the real call; see costs.

Provider keys

The other half of the bill.

Costs

What one call actually cost, line by line.

Get started

Build

Guides

Pricing and credits

The platform fee

The fee is waived on failed calls

Reading the balance

Buying credit

Running out

The ledger

Estimating cost per minute

Next

Provider keys

Costs

​The platform fee

​The fee is waived on failed calls

​Reading the balance

​The signup credit

​Buying credit

​Running out

​The ledger

​Estimating cost per minute

​Next

Provider keys

Costs

The platform fee

The fee is waived on failed calls

Reading the balance

The signup credit

Buying credit

Running out

The ledger

Estimating cost per minute

Next