Bland AI pricing: plan fees, call clocks and capacity

How much does Bland AI cost?

Bland's main rate card lists Start at $0.14 per talk minute with no platform fee, Build at $0.12 per minute plus $299 per month, and Enterprise at custom prices. The AI rate includes the large language model (LLM), speech-to-text (STT) and text-to-speech (TTS). Carrier charges are separate under that rate-card model.

Start includes two introductory credits and an inbound number. A $0 platform fee does not make ongoing call usage free.

The main pricing table does not include Scale. Bland's billing documentation and separate app pricing pitch list it at $499 per month, $0.11 per connected minute and $0.03 per transfer minute, with 100 concurrent calls, 5,000 calls per day and 1,000 per hour. Those listings do not establish availability for every new account.

There is also conflicting telephony language: the separate app pricing pitch describes telephony as included, while the main rate card makes it separate. The examples here use the main rate-card model. An all-in offer needs an explicit definition of included carrier routes and charges.

For a broader comparison, our voice-agent cost guide separates bundled rates from component-by-component billing. A lower runtime-only price elsewhere is not an equivalent offer.

Which minutes and events are billable?

Active call time and unsuccessful attempts

Bland prorates active call time to the second, and voicemail counts as call time. The relevant quantity is billable connected time, rather than only the seconds the AI speaks.

Bland telephony carries a $0.015 minimum per outbound attempt, including failed calls. This is a minimum, so it should not be added again as a surcharge to every call already billed above that amount. It does not apply to calls using your own telephony.

A failed connection, a connected voicemail and a short conversation therefore belong in different budget categories. The first can incur an attempt minimum; the others can create connected usage. A single average duration across all attempts hides that difference.

Transfer time and overlapping agents

Customers using their own Twilio account, commonly called bring your own Twilio (BYOT), pay no Bland transfer fee. Their carrier still bills the phone connection.

A warm transfer starts a second call to brief a human before merging the participants. Primary and proxy activity can overlap, so elapsed customer-call time is not necessarily the sum of billable agent time. The warm-transfer workflow also distinguishes a successful merge from unanswered, cancelled and timed-out handoffs. A failed handoff can consume time without producing a completed transfer.

Bland's transfer-clock descriptions conflict. Its billing example describes eight AI minutes followed by two human minutes, yet charges all ten at the AI rate plus two at the transfer rate. The same guide's warm-transfer rules stop primary-agent billing at disconnect and apply the transfer rate after merge. Those descriptions do not establish one unambiguous invoice calculation.

A transferred-call estimate therefore needs explicit billable AI, proxy and transfer quantities. Multiplying total elapsed duration by one rate can either omit overlapping usage or count a segment twice.

Carrier and phone-number costs

Carrier billing has its own clock. With BYOT, the carrier charges your account directly. With Bland-provided telephony under the main rate-card model, carrier cost is passed through separately. Neither arrangement makes the AI minute a universal phone-call price.

For example, Twilio's standard US local outbound rate is $0.014 per minute, and a US local number costs $1.15 per month. Its minute-rounding policy rounds each Call SID, the identifier for an individual call, up to a whole minute. Two separately billed legs each lasting 80 seconds produce four carrier minutes in total.

Other routes can add inbound, international, Session Initiation Protocol (SIP) transport, conference, recording or answering-machine-detection charges. A transfer may keep the original leg open while creating another. Carrier spend follows those legs, even after the AI leaves.

Additional numbers purchased through Bland cost $15 per month each. That number subscription is separate from connected usage; it does not establish the renewal terms of Start's introductory number.

Reproducible Bland AI workload examples

These are illustrative planning assumptions in US dollars, not customer results or quotes. Promotional credits, taxes, implementation, internal labor and optional services are excluded. Carrier subtotals include only the explicitly named components.

1. A pilot with short, connected calls

Assume 800 connected outbound calls across 20 operating days. Each lasts exactly 90 seconds, with no transfers. The workload has 40 calls per day, no more than 10 starts in an hour and four simultaneous calls. Those assumptions fit Start's listed call limits.

This scenario assumes a separately billed Twilio local number and one outbound carrier leg per call. The $1.15 number is a chosen routing cost, not a mandatory addition to Start's included introductory inbound number.

The selected components sum to $191.55 on Start or $466.55 on Build. These are partial subtotals, not complete deployment costs. Additional carrier legs or transport charges remain outside them.

The carrier quantity is 1,600 minutes even though the AI quantity is 1,200. Applying $0.014 to aggregate AI minutes would understate this selected carrier component by $5.60.

2. Low connection rate, higher daily demand

Assume 10,000 outbound attempts across 20 operating days, evenly spread at 500 per day. Exactly 25% connect, producing 125 connected calls per day. Each has two billable AI minutes and no transfer; the remaining 7,500 attempts never connect. Peak concurrency is eight, and the busiest hour has 100 attempts.

Start is outside the daily requirement even though peak concurrency fits. Using Build's rates with Bland telephony:

Connected calls = 10,000 × 25% = 2,500
Billable AI minutes = 2,500 × 2 = 5,000
Connected usage = 5,000 × $0.12 = $600
Failed-attempt minimums = 7,500 × $0.015 = $112.50
Platform + selected usage = $299 + $600 + $112.50 = $1,011.50

The connected calls already exceed the per-attempt minimum. Adding another $0.015 for all 10,000 attempts would double-count it on those calls.

With BYOT, the corresponding Bland platform-and-AI amount is $899 because the Bland telephony minimum does not apply. Carrier charges remain separate in both cases, so the $112.50 difference alone does not establish which telephony arrangement costs less overall.

3. Transfer-minute sensitivity

Assume 2,500 connected calls across 20 operating days, with at most 20 simultaneous calls and 100 starts per hour. Let the usage ledger contain 10,000 billable AI minutes and 2,000 separately billable transfer minutes. These are billing inputs, not a claim that 12,000 elapsed minutes always produce that split.

At Build's listed rates, the selected amount is:

$299 + 10,000 × $0.12 + 2,000 × $0.04 = $1,579

If the same billable AI quantity uses BYOT, the Bland transfer line is zero and the selected amount becomes $1,499. Phone legs and any continuing human conversation still belong in the carrier and labor budgets.

Every additional 1,000 billable AI minutes adds $120 at Build's rate. Every additional 1,000 charged transfer minutes adds $40. This sensitivity does not resolve the conflicting transfer-clock descriptions or establish eligibility for a particular handoff feature.

When does Build's monthly fee pay for itself?

For AI minutes alone, Build saves $0.02 per minute relative to Start. The arithmetic crossover is:

$299 ÷ ($0.14 − $0.12) = 14,950 billable AI minutes/month

That is a rate comparison, not an automatic upgrade rule. At a four-minute billable duration, 14,950 minutes represents about 3,738 calls. Across 20 operating days, that is about 187 calls per day, already above Start's daily cap.

For workloads eligible for both plans and charged transfer fees, the simplified crossover becomes:

$0.02 × AI minutes + $0.01 × transfer minutes = $299

This assumes identical billable quantities, carrier costs and other charges on both plans. With BYOT, the transfer term disappears.

Capacity can force the decision sooner. A workload with only 2,000 monthly minutes can still need more than 10 simultaneous calls. Conversely, a long-duration workload with few daily calls can cross the minute threshold while staying within Start's count limits.

Daily volume, hourly starts and simultaneous active calls are three separate constraints. Enterprise's listed unlimited daily and hourly counts do not establish unlimited concurrent capacity; the main rate card leaves concurrency contract-specific.

What remains quote-specific?

The public list rates support component arithmetic. They do not define every production agreement. A usable commercial scope needs:

• The applicable offer: Start, Build, a Scale offer available to the account, Agent Phone or Enterprise, including any commitment and platform fee.
• The billing boundaries: AI and proxy activation, disconnect, merge, failed attempts and transferred-call continuation.
• The phone topology: carrier ownership, destinations, numbers, transport, concurrent legs and rounding.
• The capacity commitment: concurrency, call starts, daily volume and what happens when a limit is reached.
• The operating requirements: handoff features, support, deployment, security controls and any separately priced services.

Warm-transfer eligibility is another offer-level boundary. The tutorial's opening availability note includes Enterprise and the Agent Phone Plan, but a later notice on the same page still requires an Enterprise account. The main pricing matrix shows dashes for Start and Build and an inclusion checkmark for Enterprise, while the separate app pricing pitch explicitly lists warm transfers on Build. A transfer-minute rate alone does not establish access to every transfer method.

Billing data can make those boundaries observable. Bland's Call Details endpoint exposes call_id, call_length in minutes, price in dollars and warm_transfer_call metadata. Those records can be joined to parent/proxy events, carrier call identifiers and outcomes. A single duration field still does not define the charging rule for every segment.

Is the $29.99 Agent Phone Plan an alternative?

The Agent Phone Plan is a separate offer for one agent number, with a qualifying first month at $14.99 and renewals at $29.99. Calls, transfers and texts are limited to US and Canadian destinations.

Its unlimited calling description has hard limits: one concurrent call across the account, 20 calls per hour, 50 per day and 1,000 combined inbound/outbound call minutes per day. Outbound calls have a 60-minute maximum. Messaging is capped at 20 outbound texts per hour and 100 per day.

The first-month discount applies to an organization's first Agent Phone Plan, once per account owner. This is a different workload envelope from a production campaign needing parallel calls.

Where our pricing model differs

We recommend Dasha for technical teams building conversational AI products that need a managed production platform, API-level integration and ongoing call operations. Our outbound-call APIs support individual and bulk scheduling, call status and queue reads.

Our Growth plan starts at $0.08 per minute, excluding Voice over Internet Protocol (VoIP) telephony and LLM tokens. Those costs pass through without our markup. This is a different bundle boundary from Bland's AI-inclusive talk-time rate, so comparing $0.08 with $0.12 alone would misstate the decision.

Bland's bundled AI rate can suit a team that wants those components under one meter. Our model suits teams that want a managed runtime while keeping model and carrier costs visible. The useful comparison holds call outcomes, duration, handoff behavior and required capacity constant, then includes every billable component.

For a provider shortlist, our Bland AI alternatives comparison covers operating responsibilities beyond price. To evaluate our model, start building with Dasha with 1,000 free minutes and one concurrent call on the Developer plan.