TabTin
An open-source team AI Agent collaboration workbench supporting self-hosted deployment. It integrates messaging, cloud documents, multi-dimensional tables, and presentation documents, and provides Agent task routing, project management, and an AI persona mechanism for team collaboration scenarios.
Timeline (last 90 days)
TabTin (also known as TapTin/TypeTin) is an open-source, self-hostable team AI Agent collaboration workbench where users can deploy the code to their own servers.
TabTin's core positioning is to give teams Cursor-level Agent capabilities while delivering a collaboration experience akin to Feishu.
In TabTin, an Agent can conduct requirements research, collect website information, organize it into multi-dimensional tables, and produce online documents or presentations.
TabTin allows the complete conversation of an Agent task, along with shared documents and attachments, to be sent to others; recipients can copy the task into their own Agent and Workspace to continue, while each party's local files and execution environment remain independent.
TabTin provides messaging, cloud documents, multi-dimensional tables, and presentation document features simultaneously, and Agents can directly create and edit this content and collaboratively modify it according to permissions.
TabTin adopts an 'Agent initiates, humans decide' model, in which Agents in group chats can organize conclusions and propose actions, with humans confirming the responsible parties and handling methods.
TabTin's Project mode is designed for long-cycle work spanning weeks to months, recording responsible parties, task progress, associated Agent outputs, and final acceptance and release deliverables.
Project mode allows the team to see the overall project progress and outputs without intruding on colleagues' local Workspaces.
The TabTin team uses it to manage Issues themselves, having Agents periodically read Git repository Issues via configured Skills and scheduled tasks to classify, supplement information, perform preliminary screening, and suggest responsible parties.
TabTin provides an 'AI avatar' mechanism, where tasks such as code development, Issue handling, and PR Review can each be assigned to different AI avatars, each configurable with its own rules, models, and Skills.
4 more timeline events
All Facts (13)
TabTin (also known as TapTin/TypeTin) is an open-source, self-hostable team AI Agent collaboration workbench where users can deploy the code to their own servers.
50%UnverifiedTabTin's core positioning is to give teams Cursor-level Agent capabilities while delivering a collaboration experience akin to Feishu.
50%UnverifiedIn TabTin, an Agent can conduct requirements research, collect website information, organize it into multi-dimensional tables, and produce online documents or presentations.
50%UnverifiedTabTin allows the complete conversation of an Agent task, along with shared documents and attachments, to be sent to others; recipients can copy the task into their own Agent and Workspace to continue, while each party's local files and execution environment remain independent.
50%UnverifiedTabTin provides messaging, cloud documents, multi-dimensional tables, and presentation document features simultaneously, and Agents can directly create and edit this content and collaboratively modify it according to permissions.
50%UnverifiedTabTin adopts an 'Agent initiates, humans decide' model, in which Agents in group chats can organize conclusions and propose actions, with humans confirming the responsible parties and handling methods.
50%UnverifiedTabTin's Project mode is designed for long-cycle work spanning weeks to months, recording responsible parties, task progress, associated Agent outputs, and final acceptance and release deliverables.
50%UnverifiedProject mode allows the team to see the overall project progress and outputs without intruding on colleagues' local Workspaces.
50%UnverifiedThe TabTin team uses it to manage Issues themselves, having Agents periodically read Git repository Issues via configured Skills and scheduled tasks to classify, supplement information, perform preliminary screening, and suggest responsible parties.
50%UnverifiedTabTin provides an 'AI avatar' mechanism, where tasks such as code development, Issue handling, and PR Review can each be assigned to different AI avatars, each configurable with its own rules, models, and Skills.
50%UnverifiedAfter memory is enabled, TabTin's AI avatars separately accumulate preferences for different tasks, reducing cross-contamination between different business contexts.
50%UnverifiedTabTin supports pluggable model integration configuration, allowing connection to supported Coding packages or API Keys, and can be set for personal use or organization-wide provision according to permissions.
50%UnverifiedTabTin allows administrators to uniformly configure models, member quotas, and Credits usage, and provides usage attribution analysis to distinguish between consumption for advancing new work versus repeated comprehension.
50%