MeetingBaaS Pricing Breakdown: What You Will Actually Pay in 2026

How MeetingBaaS pricing works

Updated October 3, 2026: an earlier version said Attendee is MIT-licensed and that MeetingBaaS's meet-teams-bot uses the Business Source License; both use the Elastic License 2.0. We also re-checked prices against each vendor's page. Details on the corrections page.

MeetingBaaS pricing has two layers: a monthly subscription that sets your limits, and token packs you purchase separately for actual usage. This makes it harder to calculate your real cost than competitors who charge a flat per-hour rate.

The subscription tiers:

Then you buy token packs on top:

Tokens never expire, which is good. But the cost per token only tells part of the story. What matters is how many tokens each meeting consumes.

What you actually pay per hour

A one-hour meeting with recording and transcription does not cost one token. Token consumption depends on which features you use:

Most teams need recording plus transcription. That means 1.05-1.25 tokens per hour, depending on whether you bring your own transcription key or use the built-in option.

The effective per-hour cost, using the most common setup (recording + Gladia transcription):

That is before the subscription fee. Add $99-299/month depending on your tier. The headline "$0.35/token" translates to $0.37-0.63/hour for recording with transcription, plus a monthly floor.

MeetingBaaS vs every alternative

Effective cost per hour of meeting recording with transcription, plus any monthly fees, taken from each vendor's pricing page (MeetingBaaS, Recall.ai, Skribby, Nylas, Attendee), checked October 3, 2026:

Recall.ai dropped prices 29% in early 2026, from $0.70/hr to $0.50/hr for recording alone (plus $0.15 for transcription). It also removed the monthly platform fee entirely. Recall.ai did not tie the cut to any competitor, but it narrowed the gap to MeetingBaaS and Skribby on the low end.

Skribby is the cheapest managed option at $0.39/hr with Whisper transcription and no monthly fee ($0.35/hr for audio with no transcription). The trade-off: a smaller company with a limited public track record, though its pricing page now carries the line "An Alianza Company", which points to a larger owner.

Nylas sits at the top of the per-hour range here, beaten only by Skribby's premium transcription models, but it bundles recording, transcription, calendar sync, summaries, and action items. Since its September 9, 2026 pricing change, every plan includes Notetaker: Free gives 5 bot hours, Essentials costs $15 a month with 10 hours and $0.80 per extra hour, and Pro costs $49 a month with 70 hours and $0.70 per extra hour ($43 and $0.65 billed annually). Existing customers who do not pick a plan move to a recommended one on December 1, 2026. If you need the full stack and do not want to assemble it yourself, the bundling has value.

Cost at scale

The cost picture shifts depending on your volume. Assuming 30-minute average meetings, recording plus transcription:

At low volume, MeetingBaaS's free tier is competitive. But if you need more than 75 bots/day or more than 2 calendar integrations, you jump to a $99+ subscription that makes the total cost 3-7x more than alternatives.

Recall.ai offers volume discounts on its Launch and Enterprise tiers, but does not publish the rates, so we cannot show what they cost. The one discounted rate Recall.ai does publish is its startup program at $0.25/hr for the first 10,000 hours, which is what the row above uses. The program is application-based. Recall.ai says the program is for early-stage teams building products where meeting data is core, but publishes no headcount or funding threshold, so there is no line you can check yourself before applying. Once you need a paid tier, MeetingBaaS's subscription overhead puts it above the simple per-hour alternatives in every scenario here.

Note on the MeetingBaaS figures: token cost is prorated to the hours used, and the Scale and Enterprise rows apply the 5% and 10% token pack discounts those tiers include.

Note on the Skribby figure: it uses their cheapest audio-only transcription model. Add $0.05/hr for video to match what Recall.ai includes as standard.

At very high volume, self-hosting is the likely cheapest option if you have the engineering team to maintain it. Attendee says self-hosting can "reduce costs by 10x compared to closed source vendors". We have not run it at this volume, so treat that as the vendor's claim.

The build vs buy question

Two source-available options (both under the Elastic License 2.0) let you skip the per-hour fees.

Attendee.dev is source-available under the Elastic License 2.0, supports Zoom, Teams, and Google Meet, and runs as a Django app in a single Docker image with Postgres and Redis, with Kubernetes recommended for production. Transcription goes through a third-party provider you configure, such as Deepgram. There is no license fee. The cost is engineering time to set it up, plus ongoing maintenance for browser automation updates when meeting platforms change their UI.

MeetingBaaS's own engine (meet-teams-bot) is also under the Elastic License 2.0. You can self-host it, but the license bars offering it to others as a hosted or managed service. The same limit applies to Attendee. It gives you MeetingBaaS's published bot engine without the token costs.

Our rough line: self-hosting makes sense when you process 5,000+ hours per month (where the managed services above cost about $2,000/month or more) and have a team that can handle infrastructure. Below that threshold, the engineering cost of maintaining a self-hosted solution usually exceeds the managed service fees.

Hidden costs to watch

Waiting time is billed. MeetingBaaS's pricing page says raw recording tokens "would be charged even when a bot has not been accepted in a meeting based on the duration the bot was waiting to be accepted." That matters more now that Microsoft Teams sends detected bots to the lobby for organizer approval by default (MC1251206). A bot that waits 20 minutes in the lobby and is never admitted still costs tokens.

Data retention. MeetingBaaS's free tier keeps recordings for 3 days. The $99 Pro tier gives you 7 days. If your workflow requires accessing recordings after a week, you need the $199 Scale tier minimum. Recall.ai gives 7 days free, then charges $0.05/hr per 30-day period for storage. At 500 hours/month kept for 90 days, that is $75/month in storage alone on Recall.ai.

Transcription is not included in the base price. On MeetingBaaS, the "starting at $0.35/token" headline assumes raw recording only. Add transcription and you consume 1.05-1.25 tokens per hour instead of 1.00. Recall.ai charges $0.15/hr extra for built-in transcription. Only Skribby and Nylas bundle transcription into their quoted prices.

Tier discounts. The Scale and Enterprise tiers include a 5% and 10% token pack discount. On those tiers the Infinity pack works out to about $0.34 or $0.32 per token.

Calendar integration limits. MeetingBaaS caps calendar integrations at 2 on the free tier, 10 on Pro. If you are building a product that connects to customer calendars, you hit the Scale tier ($199/month) fast. Recall.ai's calendar API is free with no published limits.

Bot-per-day caps. The free tier allows 75 bots per day. That sounds like a lot, but if your product serves multiple customers who each schedule meetings throughout the day, you can hit the ceiling quickly. Scaling to 300 bots/day requires the $99/month Pro plan.

Which provider fits your stage

Early-stage startup (under 100 meetings/month): MeetingBaaS's free tier or Skribby. Both work without a monthly commitment. MeetingBaaS gives you 8 free hours to start; Skribby lets you start free without a card but does not publish a trial allowance. After the trial, Skribby is cheaper per hour with no subscription overhead, at $19.50/month for 50 hours on its cheapest audio-only model, or $22 with video. Recall.ai belongs in this bracket too: teams accepted into its startup program pay $0.40/hr all-in, so $20/month for the same 50 hours, which beats Skribby at video parity and loses to it on audio only. The 10,000-hour allowance lasts years at this volume. It is application-based and the published criteria are qualitative rather than a threshold you can check, so treat it as a maybe rather than a plan.

Growing product (100-1,000 meetings/month): Recall.ai's Pay As You Go tier. No monthly fee, predictable per-hour cost, and Recall.ai says 3,000+ customers use it. The $0.65/hr all-in rate is higher than Skribby but comes with better documentation, SDKs, and support. Teams in the startup program pay $0.40/hr all-in until the 10,000 hours run out, which at 500 hours a month takes about 20 months. Recall.ai has raised $51M in total, including a $38M Series B led by Bessemer Venture Partners in September 2025 (SiliconANGLE), which suggests it has runway.

Scale (1,000-10,000 meetings/month): Negotiate with Recall.ai's Launch or Enterprise tiers. Custom pricing is available at this volume, though Recall.ai does not publish the rates, so we cannot say how far below list they land. MeetingBaaS's token model gets expensive because the subscription fee persists on top of already-high token consumption.

Privacy-sensitive or high-volume (10,000+ meetings/month): Self-host with Attendee.dev or MeetingBaaS's source-available engine. Attendee claims a 10x saving over closed-source vendors; we have not verified it. The trade-off is real: you need engineers who can maintain browser automation infrastructure, handle platform API changes, and keep transcription pipelines running.

The meeting bot API market is repricing fast. Recall.ai cut rates 29% this year. Skribby undercuts the field. MeetingBaaS's token model adds complexity that its competitors have dropped. Before you commit to any provider, run your expected volume through the actual pricing math, not the headline rate. The difference between the advertised price and your real monthly bill can be 2-3x.