← /editions
#106

Meta's Muse Crosses 1M Daily Users

// Meta's personal AI agent Muse quadrupled its daily users in about eight days, reaching 1M+ DAUs and 3M+ weekly prompters. Distribution is winning round one of the consumer agent race.

$ today_in_10_seconds

Meta's personal AI agent, Muse, now has more than 1 million daily active users and over 3 million weekly prompters, according to internal data seen by The Information. Eight days ago it had 250,000 daily users. This is the clearest sign yet of what happens when an agent ships with Meta-scale distribution.

$ lead_story

Meta's Muse Crosses 1M Daily Users

Internal data reviewed by The Information shows Muse has more than 3 million users who send it at least one prompt a week, and more than 1 million who prompt it daily. Muse's daily users roughly quadrupled in about eight days. Its daily-to-weekly ratio is around 1 in 3, which is unusually sticky for a product that's only a few weeks old. For founders building consumer agents, this changes the benchmark. The competition is no longer other startups. It's a default agent sitting inside apps billions of people already open every day.

// why_it_matters: Distribution just beat differentiation in round one of the consumer agent race.

read more →

$ also_worth_attention

  • Rewind: Muse's first week brought 500K users, 250K DAUs and 2M prompts

    In its first week, Muse drew more than 500,000 users, half of whom used it daily, and logged over 2 million prompts. That works out to about four prompts per user for the week. That figure is worth watching, because a good agent should need fewer prompts as it handles more multi-step work on its own. That makes prompt volume a shaky proxy for the value an agent delivers.

    link →

$ bigger_picture

The trend behind the headlines

Two data points eight days apart give the first real growth curve for a mass-market personal agent. The curve is steep. The caveat is that these are internal numbers counted by prompts, not by tasks completed, so they measure whether people come back, not how much work the agent actually does for them. That gap is where independent builders can still win. A general-purpose agent bundled into a social app will struggle to match deep integrations, vertical workflows and agents that reliably finish long tasks. If your agent's pitch is "an assistant you can talk to," Meta just made that free.

$ tool_of_the_day

Langfuse

If prompt counts are a weak signal, you need better ones. Langfuse is an open-source observability and analytics platform for LLM apps and agents. It traces multi-step runs and tool calls and lets you score outcomes, so you can measure tasks completed rather than messages sent.

explore Langfuse →

$ subscribe --free

Get the next edition in your inbox

// DAIly Agent covers agentic AI three times a week

// free, no spam, exit anytime