Today Personal AI

Overview

Today is a personal-agent product presented as a named, continuous digital assistant that learns a user over time, initiates useful work, and progresses from reminders toward task completion.

Current Profile

Today combines long-term personal memory with proactive service. Users can connect sources such as email, calendars, Slack, Notion, documents, and recordings; the product is described as preparing daily priorities, meeting material, research, messages, scheduling, and follow-up while learning preferences, relationships, and recurring goals.

Its intended interface is one familiar front agent rather than a visible team the user must coordinate. The product’s claimed advantage is not raw connector breadth but a context-interpretation layer that decides whose information a fact concerns, when it was true, whether it remains relevant, and whether the next step should be a quiet update, suggestion, confirmation request, or action. The team frames success as returning time and attention to the user, but the source does not establish broad reliability, retention, payment willingness, or product-market fit.

Key Characteristics

• Centers the product on continually revised memory and calibrated initiative.
• Uses connected personal and work services to prepare and coordinate recurring tasks.
• Presents one named front agent while allowing specialist execution behind the interface.
• Tries to lower repeated explanation through retained preferences, relationships, and task context.
• Gives users visibility into recorded memory plus deletion and account-exit controls.
• Seeks to progress from information and reminders toward low-risk execution with graduated confirmation.
• Favors a general personal context layer while prioritizing deeper recurring workflows.

Evidence

• Memory and initiative: 277.从提醒你，到替你办：Today想把Personal AI带到哪一步？ describes memory and proactivity as Today’s paired core capabilities and says either becomes weak without the other.
• Workflow and delegation: 277.从提醒你，到替你办：Today想把Personal AI带到哪一步？ gives scheduling, meeting preparation, research, recording, task distribution, and long-term learning or fitness support as intended uses.
• Interface and identity: 277.从提醒你，到替你办：Today想把Personal AI带到哪一步？ says users can name the agent and argues for one context-rich front agent rather than requiring users to manage multiple agents or sessions.
• Control and execution: 277.从提醒你，到替你办：Today想把Personal AI带到哪一步？ reports inspectable and deletable memory, full clearing on exit, and a path from reminders through confirmation toward low-risk action.

Qualifications

The profile rests on one launch-period podcast summary featuring the team, not an independent product audit. Task success, privacy implementation, model evaluation, user growth, retention, product-market fit, cost, and willingness to pay are not independently established. Episode 275’s Town/Today naming ambiguity does not prove that Town and Today are the same product, so this page keeps them distinct.

What Changed

• Created the product entity from its launch-period episode.
• Separated Today from the earlier Town/Today transcript ambiguity.
• Recorded identity-aware memory, low-interruption initiative, and reduced delegation friction as the product’s central thesis.

Relationships

• Personal Life Agent / 个人生活智能体 - product category organized around one continuing person across work and life.
• Personal Agent Understanding Layer / 个人Agent理解层 - interpretation layer that turns connected data into current and relevant context.
• Agent Delegation Friction / 智能体委托摩擦 - product problem Today seeks to reduce through retained background and execution.
• Personal Memory Identity Resolution / 个人记忆身份解析 - memory requirement for assigning facts to the correct person and time.
• Proactive Agents - initiative capability that decides when to prepare, ask, or act.
• Agent Permission Boundaries - control relationship for messages, scheduling, connected services, and execution.
• Town Personal AI - separate comparison entity affected by an earlier source’s Town/Today naming ambiguity.
• 乱翻书 - podcast source for the product profile.