OptimAI Agent Platform
OptimAI’s agent platform is the composition of three product primitives:
- Search gives agents context.
- Claw gives agents execution.
- Persona gives agents memory and identity.
The node network, reinforcement data layer, and OPI economy make those primitives useful beyond a single app session.
Agent Stack
Core Agent Types
| Agent type | What it does | Uses |
|---|---|---|
| Personal agent | Maintains user-approved preferences, projects, and recurring workflows. | Persona, Search, Claw |
| Research agent | Gathers current context and produces source-backed briefs. | Search, Claw |
| Workflow agent | Executes repeated tasks such as monitoring, comparison, and reporting. | Claw, Persona |
| Data agent | Creates, refreshes, validates, or packages datasets. | Claw, Data Engine, campaigns |
| Marketplace agent | A reusable specialist that can be published, discovered, and paid for. | Persona, marketplace, OPI |
Why OptimAI Agents Are Different
They are grounded in live context
Agents can use OptimAI Search and the data network instead of relying only on model memory.
They can operate, not only answer
Claw gives agents a runtime for extraction, task execution, and workflow completion from inside the Core Node environment.
They can retain user-approved memory
Persona helps agents preserve preferences, source lists, project context, and workflow decisions over time.
They can become network assets
Marketplace and OPI primitives create a path for specialized agents, datasets, and workflows to be packaged, reused, and monetized.
Example Flow
- A user asks for a market brief.
- Persona loads the user’s preferred format and watchlist.
- Search gathers current sources.
- Claw extracts companies, claims, dates, and links.
- The data layer validates and scores records.
- The agent returns a cited brief and saves approved memory.
Builder Interfaces
Builders can start with: