Questions Answered
AIM is Helium's persistent memory layer. It learns your preferences, stores context from past conversations, remembers your projects and decisions, and applies this knowledge automatically to every future interaction. You control what AIM remembers and you can delete stored memories at any time.
The Orchestrator coordinates AI models, AIM context and tools for your tasks. Results depend on the task, available context and supported capabilities; review outputs before using them.
Helium coordinates multiple AI models behind the workspace. Model availability and features can change. Contact sales to confirm the models and capabilities supported for your team; access to every model or launch-day additions is not promised.
Helium One is built by Neural Arc, the company behind ModelBeat. You can write to the team at hello@neuralarc.ai.
Our Privacy Policy states that prompts, outputs, uploaded documents and AIM memory are not used to train or improve AI models. Content is processed by the inference providers described in that policy to generate responses. Workspace permissions control sharing within your team; they do not mean processing stays entirely inside the workspace. See he2.ai/privacy-policy for details and privacy requests.
Helium One gives teams a shared AI workspace with shared knowledge, reusable prompts and administrator controls. The team plan starts at five seats. Your administrator manages member access and workspace setup.
Yes. Helium One is built for teams of five or more. Talk to the Neural Arc team at hello@neuralarc.ai about plans and additional seats.
Integrations depend on the available connectors, your workspace configuration and the permissions you grant. Contact us to confirm the tools and workflows your team needs before purchasing.
Contact the Neural Arc team at hello@neuralarc.ai to discuss your team’s needs. Once your workspace is set up, your administrator can invite members and configure shared knowledge and permissions. Existing users can sign in from the header.