OptimAI Persona Agent
Persona Agent is the memory and identity layer of OptimAI. It lets users turn approved context into a reusable agent profile: preferences, projects, trusted sources, workflows, and voice.
Search gives the agent current context. Claw gives it execution. Persona gives it continuity.
Position In The Stack
| Layer | Persona relationship |
|---|---|
| Search | Uses live context while respecting trusted sources and preferences. |
| Claw | Applies memory to research, extraction, and recurring workflows. |
| Nodes | Provide the runtime and network context behind agent tasks. |
| Reinforcement Data Network | Helps decide what context is reliable enough to reuse. |
| Marketplace | Can support reusable persona-aware agents and workflows with permission boundaries. |
Why Persona Matters
Most AI assistants reset between sessions or depend on opaque platform memory. Persona Agent is designed around explicit, user-approved memory.
That matters because useful agents need to remember:
- what the user is working on
- which sources the user trusts
- how the user prefers output
- what decisions have already been made
- which workflows repeat
- what should remain private
Memory Model
| Memory type | Examples | Default stance |
|---|---|---|
| Preferences | tone, format, risk tolerance, preferred tools | private |
| Projects | goals, plans, documents, milestones | private |
| Sources | watchlists, trusted sites, accounts, dashboards | private or scoped |
| Workflows | weekly brief, market map, report template | private |
| Decisions | prior comparisons, selected options, rejected paths | private |
| Feedback | corrections, approvals, ratings | private or aggregated |
Persona Loop
Capabilities
Personal memory
Store user-approved preferences, sources, project context, and workflow patterns.
Context-aware research
Use Search while applying the user’s watchlists, preferred formats, and known goals.
Workflow continuity
Reuse previous decisions and task structures instead of rebuilding a workflow from scratch.
Controlled sharing
Keep personal memory private by default. Share only what the user explicitly approves for a task, campaign, or marketplace agent.
Memory Governance
Persona should make memory legible. A user should be able to understand:
- what was saved
- why it was saved
- where it came from
- whether it is private, scoped, or shared
- how to edit or delete it
- which agent or workflow can use it
Example Personas
| Persona | What it tracks |
|---|---|
| Founder | competitors, hiring plans, investor updates, roadmap decisions |
| Researcher | topics, papers, citations, source quality, open questions |
| Trader | narratives, risk rules, protocols, on-chain and social signals |
| Creator | voice, audience, content pillars, campaign performance |
| Developer | stack preferences, APIs, codebase patterns, issue context |
Relationship To The Stack
- Search brings in new context.
- Claw executes research, extraction, and workflow steps.
- Persona decides what should be remembered.
- Nodes provide the distributed runtime and data layer.
- OPI can support premium agent services, marketplace use, and reward flows.
Builder References
See Persona API preview, SDK memory example, and Persona memory schema.