Free by Default: Why Meta Priced Muse the Way It Did
Muse is free for most of what people need, with paid tiers on top and transaction fees planned for later. That's a distribution strategy, not a pricing model — and the two aren't the same thing.
Of the three agents that launched or expanded this week, only one didn't ask for a subscription first. OpenAI's Dots requires ChatGPT Pro or Business Premium, with the first dot free and everything beyond that undisclosed. xAI's Grok Bot rides on SuperGrok or Cursor Pro tiers, drawing from a separate quota. Meta's Muse is free for most of what people need, with subscription plans available for people who want more, and — per Meta's own framing at Meta Connect 2026 — "free for a huge number of tokens" for now, with transaction fees planned as a future monetization path.
That's not an accident of timing or generosity. It's the same playbook Meta has run since Facebook and Instagram: remove the price barrier entirely, win the largest possible user base, and monetize the attention and activity later through a mechanism other than a subscription fee.
Why free makes sense for Meta specifically, and less sense for the other two
OpenAI and xAI both make their money directly from the product a user is holding — a ChatGPT subscription, a SuperGrok or Cursor subscription. Gating an agent behind that subscription is a straightforward extension of a business model that already depends on people paying for access. Meta's core business has never worked that way. Instagram and Facebook are free, and Meta makes its money from advertising built on top of the attention and data that free access generates. Pricing Muse the same way isn't a departure from that model — it's the model, applied to a new surface.
This also explains why free-with-limits is a much lower-risk bet for Meta than it would be for either competitor. OpenAI is spending real compute on every Dots session with no revenue if the user churns before upgrading past the free dot; that math has to work through eventual subscription conversion. Meta can subsidize a much larger free tier because it isn't relying on Muse's subscription revenue alone to justify the cost — it's relying on Muse strengthening the platforms (Instagram, Facebook, WhatsApp) that already carry the company's ad business, plus whatever the announced-but-unspecified transaction fees eventually generate.
The transaction-fee model is the part to watch, not the free tier
The free tier is the least interesting part of Muse's pricing, because it's the part Meta has already committed to and explained. The transaction-fee plan mentioned at Meta Connect 2026 is the part with real strategic weight and almost no public detail. Muse already routes purchases through Stripe's Link, with one-time-use cards for online purchases, and Muse for Small Business connects directly to Shopify and Stripe on the merchant side. A transaction fee layered on top of that flow would put Meta in a position similar to a payment processor or marketplace — taking a cut of commerce that happens to route through an agent it built, rather than charging for access to the agent itself.
Free access is the hook. A transaction fee on every purchase Muse helps you make is the business model — and Meta hasn't said what that fee looks like, or when it starts.
Meta hasn't disclosed the size of any transaction fee, which purchases it would apply to, or whether it would apply on the consumer side, the small-business side, both, or neither. Until that's public, "free by default" should be read as a launch-phase pricing decision rather than a permanent one — Meta has left itself every option, including ones that could make Muse considerably more expensive to use heavily than either Dots or Grok Bot, just charged in a way that doesn't show up as a subscription line item.
What free-by-default buys Meta competitively
Removing price as a barrier is almost certainly part of why Muse topped app charts within days of launch — a free download is a much smaller ask than committing to a ChatGPT Pro tier or a SuperGrok subscription just to try an agent. It also fits neatly with Zuckerberg's positioning of Muse as something built for everyone rather than for power users willing to pay upfront — a mass-market agent needs a mass-market price, and a mass-market price, for a company with Meta's ad-driven history, is zero.
The tradeoff is that free access at scale is expensive to run, and Meta's eventual answer to "how does this pay for itself" — a transaction fee, ad adjacency, or something not yet announced — will define whether Muse ends up looking more like a genuinely free utility or a commerce layer that takes its cut quietly, purchase by purchase, well after the free-tier headlines have faded.
What "usage limits" might actually mean
Meta's phrase — free for most of what people need, subscription plans for people who want to do more — leaves open exactly what counts as "more." For a conversational product, usage limits typically mean a cap on messages, tokens, or agent-hours per period, similar to how ChatGPT's free tier is capped relative to its Plus and Pro tiers. Given that Muse's tasks can run in the background for extended periods — negotiating, monitoring a saved goal, executing a multi-step booking — the meaningful limit for Muse may end up being less about conversation volume and more about how many concurrent long-running tasks a free user can have active at once, or how much compute-intensive background work (image generation for the avatar, real-time voice processing) is included before a paywall appears. None of that has been specified publicly, which makes it hard to know today whether "free for most of what people need" will still feel true to a small business running several Muse tasks simultaneously across Shopify, Instagram, and Slack.
The comparison that puts the risk in context
It's worth remembering that free-by-default has a real precedent for going wrong at a company Meta's size: Meta has previously scaled free products aggressively and adjusted monetization later in ways that surprised users — algorithmic feed changes, ad-load increases, and shifting data-use policies have all followed a similar "grow first, monetize later" arc on Facebook and Instagram. There's no reason to assume Muse's eventual transaction-fee model will follow that same pattern, but the historical base rate for how Meta has handled "free now, monetized later" products is a reasonable input when judging how much weight to put on today's generous pricing holding steady once the transaction-fee mechanism actually launches.
Part 17 of 30 in the Meta Muse series. Previously: Muse's App-Chart Run: What Topping the Charts Actually Signals. Next: Muse's Enterprise Platform: The Announcement Nobody Expected. Browse more builder insights or explore AI skills at aiskill.market.