Capabilities
What Health Loops can do.
The platform behind a health agent, in the terms a buyer checks: how a run is orchestrated, what it can reach, where it reaches people, what it remembers, who can stop it, and what the record shows. Every claim links to the page or doc that shows the mechanism.
The map
Fourteen capabilities. Eight pages.
The same 14 axes independent buyers and analysts use to compare agent platforms. Open the page for the mechanism, the code and the limits.
Integrations & Tool Calling Typed tools, an allowlisted fetch, health tools and Shopify.
Workflow Orchestration Checkpointed steps, retries, waits and versioned deploys.
Knowledge Grounding & RAG Your PDFs, text and markdown, searched by the agent.
Human Oversight & Guardrails Approvals, tool switches, host allowlists and handoff.
Security, Identity & Governance Own store, scoped keys, roles, secrets, POPIA and GDPR.
Observability & Auditability Every run, step, tool call and log line on one timeline.
Memory & State Persistence Threads, a shared user record and durable run state.
Deployment & Data Residency Managed runtime with your own data plane and DPA terms.
Prebuilt Agents, Templates & Packs Nine ready-made agents with vertical variants.
Triggers & Channel Coverage API, dashboard, store events, WhatsApp, SMS, email, phone.
Model Flexibility & Routing A model and reasoning depth chosen per agent.
APIs, SDKs & MCP Extensibility HTTP trigger, TypeScript SDK and the health CLI.
Testing, Debugging & Optimization Playground, versions, runs, steps and event logs.
Browser & Computer Use Not a fit. Loops acts through APIs, tools and channels, not a browser.
By page
Start with the question you have.
Workflows and orchestration
Durable steps, waits and retries.
Integrations and tools
Your systems, health tools and Shopify.
Triggers and channels
API, store events, WhatsApp, SMS, email, phone.
Knowledge and memory
Your documents, and what the agent remembers.
Human oversight
Approvals, consent and handoff.
Security and audit
Isolation, keys, data and the run record.
APIs, SDKs and models
One POST, TypeScript, the CLI and model choice.
Prebuilt agents
Nine ready-made jobs, with vertical variants.
For buyers and analysts
The evidence, not the adjective.
Every page links to the first-party source for its mechanism: the developer docs for the runtime, security page for the data terms, pricing for the entry point, and the agent directory for the templates. Where a capability is not offered, the page says so.
Ollie’s agents give wellness guidance. They don’t diagnose or prescribe, and they hand off to a human expert when it matters.
Bring the journey. We will show you the loop.
