Some work starts privately. Some work needs collaboration. TeamUp supports both. Users can keep personal chats organised in projects, work inside shared team spaces, invite teammates by email, and move between individual thinking and shared discussion without switching tools.

Because messaging, team chats, and shared workspaces live in the same product, collaboration stays closer to the work itself. That makes it easier to keep context, decisions, and next steps connected.

TeamUp is a Bring Your Own Key (BYOK)-first AI workspace. Users are expected to connect their own API keys for supported providers, giving them the flexibility to choose the models that fit their work instead of being locked into a single AI setup.

That approach gives users more control over both workflow and cost. If a person or team already has access to preferred providers, they can use that access inside TeamUp while keeping projects, chats, files, team spaces, and collaboration in one organised workspace — without relying on a large flat subscription just to use AI.

For users who do not want to bring their own keys, TeamUp also includes managed options in beta. The platform currently provides a 1,000,000-token allowance on a standard chat model (currently DeepSeek) and offers credits for Google Vertex - Gemini multimodal chat, image and video models inside the app.

TeamUp is deployed on Google Cloud Platform and uses Cloud KMS (Key Management Service) technology to securely store and manage users' API keys. That means your credentials are protected with enterprise-grade security, encryption, and access controls built into the platform infrastructure.

TeamUp Assistants help users generate more reliable answers and content by grounding AI responses in the documents they upload. Instead of relying only on general model knowledge, Assistants use user-provided data to answer questions with stronger relevance, better context, and more useful output.

Users can create their own Assistants by defining a role, writing a task description, and uploading the documents the Assistant should use. Once set up, they can ask questions in natural language and receive grounded answers that often include citations to the retrieved source material.

This makes Assistants especially useful for work that depends on accurate internal context — such as manuals, policies, knowledge bases, PDFs, process documents, and other structured or unstructured sources. Newly ingested content can be reflected in answers, helping users work with information that stays current inside their own Assistant.

Built on Google’s RAG Engine, TeamUp Assistants prepare uploaded content for semantic retrieval and grounded answering. By retrieving relevant context before generation, Assistants are designed to reduce hallucinations and produce answers that stay closer to the source material provided by the user.

Users can also share their Assistants with teammates, making it easier to reuse trusted knowledge across a team. Once shared, an Assistant can be used in team chats and in individual projects outside team conversations, so the same grounded knowledge stays available wherever the work happens.

This helps teams stay aligned, avoid repeating setup, and keep answers consistent across shared and personal workflows. It is a simple way to turn one Assistant into a reusable team asset.

TeamUp is live in beta with practical core workflows ready to use today. Built-in feedback through Pulse lets users report issues, suggest improvements, and help shape what comes next directly from inside the workspace.

The current focus is to stabilise the product through real user feedback, strengthen the core experience, and continue development based on user requests and planned improvements. TeamUp is being refined in the open, with each round of feedback helping move the product closer to a stronger post-beta release.

TeamUp gives users one place to organise projects, use AI, work with files, message teammates, and collaborate in shared spaces — with BYOK-first flexibility at the core.