ABOUT THE PRODUCT
ModelCrew is a multi-model AI workspace for creating role-defined AI teammates and coordinating their work on projects. The teammate keeps its role and instructions in the application; an underlying supported AI model supplies its responses. A Project Lead can delegate, review and synthesize contributions in enabled workflows.
- Product category: multi-model AI team workspace
- Primary idea: the teammate is the role; the model is selectable
- Stage: private beta / evolving availability
ModelCrew at a glance
| Question | Answer |
|---|---|
| What is it? | A multi-model AI workspace and coordination layer. |
| Who is it for? | People working on recurring projects that benefit from distinct AI roles. |
| Does it build its own LLMs? | No. It uses underlying models from external AI providers. |
| What persists? | Teammate role configuration and project information within the application. |
| Can models be changed? | A supported teammate model can be reassigned without recreating its role configuration. |
| Does it replace native ChatGPT or Claude? | No. Native consumer-app features are separate. |
| What is the access model? | Private beta; initial launch direction is ModelCrew-managed provider access and credits. |
What ModelCrew helps users do
- Define a project and give teammates distinct names, responsibilities and working instructions.
- Assign supported AI models to specific teammates, subject to model availability.
- Ask an individual teammate for an answer or coordinate multi-step work through a Project Lead where enabled.
- Supply project context to separate model requests so teammates can work from relevant shared information.
- Review contributions, approve or redirect continued work, and receive synthesized outputs.
- Keep a record of project conversation and supported files and outputs in the workspace.
What makes its approach different?
Many tools help people compare several AI models, run a single chatbot or build agents in code. ModelCrew emphasizes a project-centered working team: users define the jobs first, then choose the model behind each job.
That makes model choice flexible without confusing the teammate with its provider. It is a design approach, not a promise that every model performs equally well or that ModelCrew has unique capabilities no other product offers.
What ModelCrew is not
- Not a new foundation model or a training service.
- Not a direct connection between separate ChatGPT, Claude or Gemini consumer chat accounts.
- Not unlimited long-term model memory; project context is managed by the application and subject to model limits.
- Not a guarantee of autonomous correctness or of lower cost than a single-model workflow.
- Not a substitute for professional judgment, source verification or appropriate data security controls.
Models, providers and beta availability
ModelCrew is provider-independent in design and maintains a model catalog. However, individual providers may have different feature support. A catalog listing, a successful direct model call and full delegated Crew orchestration are not equivalent.
In the current development plan, OpenAI and Anthropic are the primary confirmed collaboration focus. Other providers are being assessed or enabled selectively. The public product page should reflect only capabilities actually available to the relevant users.
Access, credits and API credentials
ModelCrew is preparing a platform-managed API-access experience funded through ModelCrew credits. Users will not be required to connect their personal provider API keys under the initial beta/launch direction. The beta is invitation-based, and commercial plan details should be published only when finalized.
Multi-agent work can require multiple model calls, so usage depends on the models selected and the work performed. No blanket savings or unlimited-usage promise is made.
Current status and product information
ModelCrew is an evolving private-beta product. Features, supported models, pricing, access conditions and available tool integrations may change as testing continues.
For accurate information, use the current public website and in-product availability indicators. This page is intended to be a clear product overview, not a substitute for terms, privacy disclosures or the actual capabilities enabled in your account.
Frequently asked questions
Is ModelCrew an AI model?
No. It is an application and orchestration workspace that uses models offered by other providers.
Does ModelCrew offer BYOK?
Bring-your-own API keys are not planned for the initial beta/launch. The intended access model is ModelCrew-managed provider credentials with credits.
Can teammates switch models?
The role configuration is separate from the assigned model, allowing reassignment to an enabled compatible model without recreating that teammate.
Is the beta publicly available?
ModelCrew is currently in private beta, with features and access controlled during testing.