Free lets you test Cortex on a real scope. Pro unlocks advanced individual use with more topics, connectors, insights and sprints. Team adds shared memory, team agents, roles and a dedicated Cortex Cloud.
Every plan gives access to the foundations of Cortex. Cortex does not charge per message when you use your own providers or a local model.
To install Cortex on an initial scope and test it on a real project.
To use Cortex daily across multiple projects and delegate more work to it.
External sessions may come from ChatGPT.com, Claude.ai, Claude Code, Codex, Antigravity and compatible tools that record conversations on your machine.
To share Cortex across multiple collaborators, projects or clients.
Cloud volumes scale with the team. Intensive usage or dedicated infrastructure may require a specific arrangement.
| Capability | Free | Pro | Team |
|---|---|---|---|
| Application desktop | Included | Included | Included |
| Fonctionnement multi-provider | Included | Included | Included |
| Local models | Included | Included | Included |
| Structured local memory | Included | Included | Included |
| Sources & provenance | Included | Included | Included |
| Workspaces | 1 | 5 | No fixed limit |
| Active topics | 3 | 50 | No fixed limit |
| Specialized agents | 1 | 10 | No fixed limit |
| Insight history | Included | Included | Included |
| Insight generation | Non | Included | Included |
| Cross-project insights | Non | Included | Included |
| Active memory sprints | 1 | 10 | No fixed limit |
| Active team sprints | Non | 2 | No fixed limit |
| Secure connectors | 1 | 5 | No fixed limit |
| External session captures | Non | 100 / month | No fixed limit |
| Skills.Factory imports & exports | 1 import + 1 export | No fixed limit | Synchronization |
| Dedicated Cortex Cloud | Non | Non | Included |
| Shared memory | Non | Non | Included |
| Roles & permissions | Non | Non | Included |
| Team audit log | Non | Non | Included |
« No fixed limit » signifie qu'aucun quota standard ne bloque l'usage normal du plan. Les fonctions cloud restent soumises à une politique d'usage raisonnable et au dimensionnement de l'infrastructure.
Reaching a quota blocks new costly actions. Existing topics, sources and stored information remain accessible.
Cortex indicates the affected capability, the limit reached and available options: wait for renewal, free up a resource or change plans.
An already-produced action doesn't disappear because a quota is reached between its creation and validation.
After a downgrade, data remains. Only new writes reserved for the higher plan are suspended.
Provider subscriptions and usage costs are not included in the Cortex price. You can use your compatible subscriptions, your own API keys, local models and multiple providers per mission. This separation lets you switch models without moving your project memory.
No. Free is designed to remain usable indefinitely on a reduced scope.
Yes. Local memory and local models can be used without Cortex Cloud. Some synchronization, collaboration and automation features require remote services.
You keep working in your tools. Cortex connects the conversations, documents, decisions and actions you choose to entrust to it.
Yes. Project memory stays in Cortex. It doesn't depend on the model used for a particular conversation.
Data is not deleted. Paid capabilities become read-only or stop accepting new actions until the plan is reactivated.
No. Selected participants test features matching their use case for free during the beta.
The private beta tests memory, connectors, captures and multi-agent sprints against real, long-running projects. Access opens gradually to verify the product matches each participant's needs.
