On this page
Character.AI reported more than 20 million monthly active users in February 2025. That is a dated company disclosure, not a verified live user count for today. Web visits, app downloads, messages, and registered accounts describe different things and should not be presented as interchangeable user statistics.
This reference separates verifiable company statements from figures we could not substantiate. It also explains what each number can tell a researcher or marketer.
Verified Character.AI statistics
| Metric | Reported figure | Date and source | Scope |
|---|---|---|---|
| Monthly active users | More than 20 million | February 7, 2025 engineering post | Across the company’s services |
| Worldwide users | About 20 million | January 13, 2026 technical post | Described as users; not explicitly a new monthly-active measurement |
| Messages sent since launch | More than 2 billion | March 23, 2023 announcement | Cumulative messages over the period described |
| AI characters created | More than 1 million | Same March 2023 announcement | Created characters, not people |
| Voices created | More than 1 million | June 2024 Character Calls announcement | Created voices, not unique creators |
Swipe the table to see all columns.
These are first-party reports. They have not been independently audited by BrandWell. Keep the date and metric definition when quoting a row; do not relabel all the figures as current 2026 totals.
How many people use Character.AI?
The clearest monthly-active figure in the sources reviewed is the February 2025 disclosure of more than 20 million users across its services. The later technical post gives a similar overall user figure with different wording. It does not establish that growth stopped, nor does it supply a precise updated monthly-active total.
A company can report registered users, monthly active users, daily active users, or another audience definition. Before comparing two announcements, confirm that their populations and time windows match. A larger cumulative account count cannot be used to calculate growth in monthly active users.
Visits are not unique users
A person can return to a website many times in a month. A traffic-estimation service may cover the website without capturing all app usage. Its modeled visits therefore cannot be substituted for the company’s active-user figure.
For any third-party estimate, retain the provider, domain, month, geography, device coverage, and metric. Avoid mixing a website visit estimate from one month with an app download figure from another and presenting them as a single audience snapshot.
Revenue, valuation, and funding need separate labels
This review did not verify a current audited annual revenue figure from the company’s public disclosures. The previous article listed inconsistent revenue totals without sufficient primary support, so those figures are no longer presented as facts here.
Revenue measures business activity over a period. Funding is capital raised, and valuation is an assessment of company value at a particular transaction or point in time. None can substitute for the others. The March 2023 Series A announcement is a historical financing milestone; it should not be described as a current valuation update.
What about demographics and Reddit activity?
We have not verified a current platform-wide age, gender, or country distribution from the primary sources reviewed. The earlier article included a gender table adding to 110%, which was invalid. It also mixed subreddit membership and activity snapshots with platform usage.
A subreddit member count represents that community, not everyone using Character.AI. A live “online” count is a snapshot. Neither proves platform-wide daily active users or demographic composition.
How to use these statistics in a content plan
Choose the question before choosing the largest number. A market-size discussion needs a well-defined audience figure. An engagement discussion needs a time window and usage definition. A revenue analysis needs financial evidence rather than a traffic estimate.
When preparing a sourced article in BrandWell, put the metric, date, source URL, and limitation into the RankWell brief. Review the finished draft and its search preview to ensure the headline does not promise “latest” numbers that the evidence does not provide.
Correction note: This revision removes inconsistent revenue and demographic claims, mismatched traffic comparisons, and an incorrect duration comparison. It preserves the article URL while replacing unsupported figures with dated, traceable evidence.
Reviewed and updated October 2, 2026.



Farnaz Kia
Justin McGill