AI assistants are starting to show up where many people already communicate: in text messages. Instead of opening another app, users can send a request and ask an agent to handle tasks such as managing reminders, organizing calendars, researching trips, or making reservations.
One message can connect scattered information
Many of these services aim to make information from messages, email, calendars, and other connected apps easier to act on. Caddy, for example, works in iMessage for iPhone users and RCS for Android users. It can identify details that may need follow-up, add events to calendars, set reminders, track follow-ups, and do research. Caddy has been available in public beta since April 2026.
Other general assistants offer a wider mix of ways to communicate and work. Comma can operate across a user’s browser, computer, and files, while staying accessible through Signal, Telegram, and WeChat. It can retain context across devices and conversations, surface completed work for review when approval is needed, and provide daily briefings or recorded-call summaries. Comma is free to start and open source.
Folk is available through iMessage, WhatsApp, and Telegram. It can remember personal context, manage reminders, research topics, track flights, handle email, and make restaurant reservations. It runs on its own private cloud computer and can execute code and multi-step tasks. Its beta launched in May 2026; it is free, with a Pro subscription for $8.33/month for unlimited background tasks.
Other services emphasize delegation and follow-through. Iris lets users hand off tasks through iMessage and is designed to follow up proactively. Martin can manage calendars, email, reminders, tasks, notes, and communications, and can send messages or make calls. Its subscriptions start at $21 per month.
Household coordination is a major use
Several agents focus on the many details families manage across school messages, activities, meals, calendars, and household tasks. Fambot works through an app and text messages, and sends a nightly summary of the following day, including events and details such as school uniform or packing requirements. Parents can reply to ask questions or update calendars. It connects to Gmail, Google Calendar, Outlook, and WhatsApp, with Apple Calendar support planned for the future.
Fambot launched in beta in early September 2026 and has raised $3.5 million in pre-seed funding. It is free during the beta period, and expects to eventually charge roughly the equivalent of a Netflix subscription.
Ohai can turn messages, forwarded emails, and voice requests into schedules, reminders, and plans. It also coordinates calendars, chores, and meals among family members. Its basic option is free, with subscriptions based on family size starting at $9.99 per month. Ollie also brings school messages, calendars, task lists, and group chats together. Launched in June 2026, it has a free tier and paid plans starting at $25 per month for 150 messages. A $100-per-month plan includes up to 1,000 messages. Ollie says it is one of the first mainstream family-focused AI assistants to achieve SOC 2 compliance.
Orbits can bring household calendars, lists, and conversations together, and help with tasks such as booking restaurants, changing appointments, requesting service quotes, and reminding family members about chores. Tomo can join group conversations and help with scheduling, goals, reminders, calendars, and email. Its basic plan starts at $19.99 per month.
Specialists take on travel, work, and creative tasks
Some agents are built around a particular setting. Miso focuses on travel through iMessage, combining AI trip planning with support from a dedicated travel team. It can arrange flights while taking account of travel preferences and loyalty points, and provide trip details, flight updates, and personalized recommendations.
For professional tasks, Town gives users a “Townie” designed to learn how they work and handle recurring tasks across email, calendars, documents, Slack, and other platforms. In June 2026, the company raised a $55 million Series A led by Andreessen Horowitz and Forerunner. Stanley, meanwhile, is aimed at content creators. It can use Instagram context to suggest ideas, turn voice notes into scripts and captions, manage a content calendar, and schedule posts.
Skye takes a different approach on iPhone, presenting contextual cards, suggestions, and briefings from connected services rather than waiting for users to open an app. It can prepare information ahead of meetings and surface reminders. Signull Labs raised about $3.6 million in seed funding in 2025; Skye entered private testing and beta in April 2026 and has not announced a public release date.
More autonomy brings new questions
Some agents can take actions beyond organizing information. Instinct supports text and voice, connects to apps and services such as email, calendar, and Google Workspace, and has been used for tasks including planning trips and buying groceries. It raised $350 million at a $2.5 billion valuation, then raised $1 billion in September 2026 at a $10 billion valuation. The company remains in private beta.
Instinct has begun giving users dedicated email addresses for assistants to use when signing up for services or contacting businesses. That can keep activity out of a user’s personal inbox, but the source also notes concerns about privacy and security. Wajo’s consumer agent, Fo, has its own email address, phone number, and payment card to interact with businesses without requiring users to share their credentials. Wajo says it can bring in a human assistant when Fo cannot complete a task. Fo launched publicly in September 2026 with a free tier.
Privacy is also central to Underdog, whose invite-only beta launched in early October 2026. Its model runs on the user’s device, and its initial release supports Mac and Windows computers. Linux, iPhone, and Android versions are planned. The company plans to offer the app for free initially and says it will never include advertising.
These agents differ in where they work, what they can access, and how much they can do without a person stepping in. For users, the practical choice is less about a single kind of assistant and more about which tasks they want to delegate and which services they are comfortable connecting.