Persona is a persistent AI runtime designed to behave more like a product-native operator than a disposable prompt session.
This documentation is written for teams evaluating, buying, and integrating Persona into real products. It does not try to mirror the full internal engineering tree.
The first product concepts to learn
Before looking at capabilities or deployment details, understand these ideas:
- a space is the top-level environment and authority boundary
- an app is a product integration identity inside that space
- a Persona is the persistent AI runtime living in that space
- runtime access happens through scoped, time-bounded runtime tokens
Who this is for
- Product teams that want an AI persona with memory, planning, and action continuity.
- Technical buyers evaluating whether Persona can fit an existing product stack.
- Integrators who need to understand the built-in surfaces, the integration model, and the deployment shape.
What Persona gives you
- A persistent brain that can remember, plan, and act over time.
- Built-in capabilities such as chat, agenda, notifications, whiteboard, document library, KV datastore, and customer billing.
- A model where those capabilities are connected directly to the Persona brain, instead of being bolted on as separate tools.
- A deployment shape that can extend into business systems, product surfaces, and embedded environments.
Best reading path
- What Persona Is
- Spaces, Apps, and Personas
- Built-in Capabilities
- Customer Billing
- Authentication and Runtime Tokens
- Integrating Business Systems
- Embedded and Simulated Environments
- Deployment Model
- Launch Checklist
What is intentionally not here
- Internal milestone tracking
- Delivery closure notes
- Repository-wide engineering history
- Low-level implementation records that are useful to core developers but irrelevant to product integration
If a public customer or integrator needs a concept, it belongs here in product language. If it only matters to internal delivery, it stays out.