Voice AI API pricing: minutes, characters and commitments.

These are the current API rates and commitments.

Frequency pricing catalog, effective 24 October 2026. All prices exclude GST.

Voice-agent API

Agent rates, inclusions and add-ons

Plan	Rate	Monthly commitment	Inclusions and add-ons
T1	₹1/minute	₹1,00,000	All models and features. Single-prompt and node-based agents. Bring your own telco, WhatsApp or browser calling. Optional Mirai phone calling: ₹0.30/min. Optional call insights: ₹0.20/min; enable in Billing
T3	₹3/minute	₹50,000	All models and features. Single-prompt and node-based agents. Bring your own telco, WhatsApp or browser calling. Mirai phone calling and call insights included
T5	₹5/minute	None — pay as you go	All models and features. Single-prompt and node-based agents. Bring your own telco, WhatsApp or browser calling. Mirai phone calling and call insights included

Agent duration is rounded up in 10-second blocks, with a one-block minimum. For a billable call:

cost = ceil(duration_seconds / 10) × 10 / 60 × rate_per_minute

Base agent charges before GST; T1 telco and insights are additional

Billable call duration	Billed duration	T1	T3
8 seconds	10 seconds	₹0.17	₹0.50
59 seconds	60 seconds	₹1	₹3
61 seconds	70 seconds	₹1.17	₹3.50

Every company starts on T5 at ₹5/minute without a commitment. All tiers have the same models and features. T1 offers optional Mirai phone calling at +₹0.30/minute and insights at +₹0.20/minute. BYO telco, WhatsApp and browser calls have no Mirai telco charge.

Standalone Fourier TTS

₹1.80 per 1,000 characters. Shared endpoint. No contractual SLA.

Shared-endpoint TTS plans

Plan	Monthly commitment	Concurrent requests
Pay as you go	None	Up to 5
Business	₹50,000	Up to 25

Monthly commitments are credited to your wallet. Unused credit never expires. Top up anytime at the same rate.

Enterprise: Custom volume pricing, concurrency, dedicated endpoint and SLA. Discuss your workload.

How is a call billed?

Each connected call rounds up to the next 10 seconds, one block minimum, at your plan’s rate: 59 seconds bills as 60, and 61 as 70. The call’s cost field and your wallet ledger are the real record.

You pay only when the line connected:

Outcome	status	Billed
Conversation finished	completed	Yes
Answering machine	voicemail	Yes
Hit the duration cap	timeout	Yes
Rang out	no_answer	No
Line busy	busy	No
Couldn’t connect, or our pipeline errored	failed	No
You cancelled it	aborted	No

Before dialling, we check that your wallet covers one minute at your rate. If not, POST /v2/calls returns 402 insufficient_balance and no call is created. Calls already running finish. A campaign that runs low pauses with pause_reason: "insufficient_balance" and resumes after you top up.

Realtime sessions bill the same way, from when the audio goes live until the socket closes. There’s no charge per token, request or tool call.

How much will my calls cost in a month?

Round each answered call to whole blocks first, then multiply. Averaging the seconds first comes in a little low. The docs give this formula:

monthly ₹ = answered_calls × mean(ceil(secs / 10)) × rate / 6

Say you dial 10,000 people a month, 40% answer, and calls average 75 seconds. Before GST:

10,000 × 0.40 = 4,000 answered calls
75 seconds -> 8 blocks = 80 billed seconds
T5: 80 / 60 × ₹5 = ₹6.67 a call (rounded up to the paisa) × 4,000 = ₹26,680
T3: 80 / 60 × ₹3 = ₹4.00 a call × 4,000 = ₹16,000 of usage, below the ₹50,000 minimum

The 6,000 unanswered calls are free. T3 would still bill its ₹50,000 minimum, so T5 wins here.

When does a commitment plan pay off?

When you call enough to fill it. T3 and T1 are minimum-spend plans, paid upfront into your wallet. Paid calls count towards the minimum, and at month end we settle only the gap. Promotional credit doesn’t count, and extra top-ups carry forward.

T5 vs T3: ₹50,000 / ₹5 = 10,000 billed minutes a month
T3 vs T1: ₹1,00,000 / ₹3 ≈ 33,334 billed minutes a month

Below 10,000 billed minutes a month, stay on T5. On T1, phone calling and insights are extras that T3 includes, so add them if you’ll use them. Your company sits on one plan for every call. To cap one campaign’s spend, set budget_paise; the campaign guide shows how.

How much will text to speech cost?

Count the characters you’ll actually send, rather than guessing from call minutes. 1,000 replies of 120 characters is 120,000 characters, or ₹216.

Each request rounds up to the paisa. A 59-character request we sent on 16 October 2026 came back with X-Cost-Paise: 11, because 59 × ₹1.80 / 1,000 = ₹0.1062. Failed, stalled or abandoned audio is free, and format and sample rate don’t change the price. On the WebSocket, each sentence bills on its own and paisa fractions carry over.

Concurrency counts requests in flight at once, per workspace. A request over the limit gets a free 429 at_capacity with Retry-After; ask us if your limit doesn’t match your plan. Log X-Request-Id, X-Chars-Billed and X-Cost-Paise to check your forecast.

What does transcription cost?

₹18 per hour of audio, on every plan. POST /v2/transcriptions charges a job once, when it completes, for the recording’s length to the millisecond, rounded up to the paisa. A five-minute call costs ₹1.50. Degraded and failed jobs are free. The Hindi speech-to-text guide covers the API.

What should I check before going live?

Prices exclude GST, and an existing agreement keeps its terms. A pilot workspace runs 5 calls at once, and queues up to 500 more before new requests get 429. We can raise both.

Until your company goes live, calls run at T5 on the sandbox’s ₹200 credit. Go live in the console is self-serve and starts you on T5, with no commitment. For a dedicated endpoint, custom concurrency or a contractual SLA, ask about enterprise terms.

Next, compare integrated and assembled stacks on cost, or start with the voice API overview or Fourier TTS setup.