Bland AI Pricing Explained: Plans, Caps and Extra Fees

Bland AI pricing is one per-minute rate that includes its own models: $0.14 on Start with no monthly fee, $0.12 on Build plus $299 a month, and custom on Enterprise. Phone time, outbound attempts, transfer minutes and texts are billed on top. Dograh, an open-source, self-hostable voice agent platform, has no per-minute platform fee when self-hosted.

Key Takeaways

• Bland's rate includes its own models; phone time, attempts and transfers are billed on top.
• Start's 100-calls-a-day cap often forces Build before the 14,950-minute break-even does.
• To pick your own models and skip a per-minute platform fee, self-host Dograh.

This post is part of our guide to Bland AI alternatives. The guide's section "What Bland AI really costs" gives one worked bill; this post prices every line Bland can charge, plan by plan, with the caps that decide when you move up.

Bland's plans at a glance

Bland sells two self-serve plans on its pricing page, a third on its billing page, and Enterprise by contract.

Start needs no card and comes with two credits and an inbound number, which Bland values at $15 a month. Build adds five voices and 50 knowledge bases. Enterprise adds the business associate agreement (BAA) for HIPAA, single sign-on, data residency and on-prem or virtual private cloud (VPC) deployment. A separate Agent Phone plan, $29.99 a month after a $14.99 first month, gives one agent a single US number with unlimited calls and texts in the US and Canada, capped at one call at a time and 50 calls a day.

What the per-minute rate includes

Bland's rate covers its AI models, so you never see a token bill for a call. The phone line sits outside it.

Bland runs its own speech and language models, and its pricing page promises "No token charges. No model-provider pass-throughs." Call time is billed to the second, only while the call is active, and voicemail counts as normal call time. The trade-off is choice: on Bland's self-serve plans you cannot move to a cheaper language model or another voice provider, so the plan rate is the only lever on that part of the bill.

The rate has moved recently. On 5 December 2025, Bland switched self-serve customers from a previous standard rate of $0.09 a minute to plan-based rates, so Start went up by about 56% and Build and Scale by less. Existing customers got a one-time credit worked out from their previous 30 days of usage.

Bland's homepage says one rate covers telephony too, while its pricing and billing pages bill phone time separately, on your own carrier or Bland's Twilio at pass-through cost. That is where "predictable" needs care: the AI rate is predictable, and the bill is that rate plus everything below.

The lines on top of the rate

Phone time, outbound attempts, transfer minutes and texts each add their own line.

Phone time. You bring your own carrier, or use Bland's Twilio at pass-through cost. Bland sells numbers in the US and Canada at $15 a month. If your carrier is Twilio, its US list price is $0.014 a minute for outbound calls and $0.0085 for inbound calls on a local number, so 10,000 outbound minutes add about $140. Connecting your own carrier over SIP (Session Initiation Protocol) needs an entitlement that Bland support switches on for Enterprise organisations, and those calls pay Bland's plan rate plus the carrier's charge.

Outbound attempts and failed calls. On Bland's telephony, each outbound attempt and each failed call carries a $0.015 minimum charge. For a list with low answer rates this adds up: if 10,000 dials go unanswered, the minimum alone comes to $150.

Transfer minutes. After a call is transferred to a person on a Bland number, the transferred time is billed at $0.05, $0.04 or $0.03 a minute by plan. On your own Twilio account, transfer time is free. In Bland's Enterprise transfers, where a second agent briefs the person first, that second agent's talk time is billed at the plan rate too.

Texts and chat. Texts cost $0.02 each, sent or received, and web widget messages $0.01 each. Norm, the assistant that builds agents from plain language, is billed separately by tokens.

Open Source Alternative to Vapi / Retell

Self-hosted voice agent platform — no per-minute fees

dograh-hq/dograh

Star on GitHub

What 10,000 minutes a month costs

At 10,000 minutes, Build is the cheapest plan that fits, once the caps are counted. Every figure here is our arithmetic from Bland's list prices, before phone time.

Start comes to $1,400 (10,000 x $0.14). Build comes to $1,499 ($299 + 10,000 x $0.12), and Scale to $1,599 ($499 + 10,000 x $0.11). Start looks cheapest, but its cap of 100 calls a day allows about 3,000 calls a month, so 10,000 minutes only fits if calls average more than three minutes.

The plans cross at two points. Build costs less than Start once you pass 14,950 minutes a month, since its $299 fee is repaid at $0.02 a minute. Scale costs less than Build above 20,000 minutes, since its extra $200 is repaid at $0.01 a minute. At 50,000 minutes, Build is $6,299 and Scale $5,999.

Add phone time, attempt minimums, transfer minutes and texts to whichever plan you pick. Our Retell AI vs Bland AI comparison uses the same 10,000-minute example, so the two posts can be read together.

When the caps decide the plan

For many teams, the call caps force a plan change before the price does.

Start allows 10 calls at once, 100 a day and 100 an hour. Build allows 50 at once and 2,000 a day, and Scale 100 at once and 5,000 a day, both with 1,000 an hour. Bland's billing page says calls over a cap are rejected rather than charged. An outbound team dialling 500 numbers a morning outgrows Start on its first day, whatever its minute count.

Enterprise removes the caps and sizes concurrency to your volume, under a custom contract. Bland's billing page points teams beyond Scale to sales.

What the same minutes cost on Dograh

On Dograh, the platform line shrinks or disappears, and the model and phone lines become yours to choose.

Self-hosting Dograh uses Docker Compose on a server with at least 8 GB of RAM and 4 vCPUs. Self-hosted Dograh has no per-minute platform fee (servers, model keys and phone minutes still cost); Dograh Cloud has a 1¢-a-minute platform fee and includes 10 concurrent calls, with more on request.

Unlike Bland, neither includes the AI models. You set them in model configurations: Dograh's own models, any supported provider on your own key, or open-weight models on your own GPU. The phone line stays with your carrier too. Dograh connects to Twilio, Vonage, Plivo, Telnyx, Cloudonix, Exotel, Vobiz or Asterisk ARI, and on Dograh Cloud, Dograh Managed SIP connects your own SIP trunk or phone system (PBX). Dograh doesn't sell numbers, so outbound Campaigns dial through your carrier at its own rates.

Dograh can also cut the voice line itself. Its agents can play pre-recorded human clips and use text-to-speech only when no clip fits, which in our experience cuts cost by roughly 3x against full text-to-speech; our tutorial on mixing pre-recorded audio with TTS shows how.

The limits are worth stating. Self-hosting means your team runs the servers and the updates, and model and phone bills arrive from each provider rather than as one line. Transfer Call to a person works on Twilio only and passes no summary of the call.

If you are pricing the other hosted platforms too, our breakdowns of Vapi pricing and Retell AI pricing use the same line-by-line approach.

Dograh is the strongest fit for a team that wants to own each line of the bill, because every one of them becomes a choice. It is open source under BSD-2-Clause, so you can self-host it without a sales contract, and self-hosted it has no per-minute platform fee. You choose every model, including open-weight ones on your own servers, and your numbers stay with your carrier. A team used to Bland's pathways gets the same nodes-and-pathways builder, with a QA node that scores each step. To try it, run it from GitHub with Docker Compose, or start on Dograh Cloud; Dograh's pricing page lists both.