Developers
Agents should understand PrimAI without scraping it.
PrimAI exposes machine-readable files so developers, agents and engines can understand the guides, the limits and the public premium API.
- ✓Neutral
- ✓No commissions
- ✓Swiss app
PrimAI developer information: OpenAPI, WebMCP, llms.txt, ai.txt, machine files and rules for agents around Swiss health premiums.
Machine readable
Clear endpoints instead of screen-scraping.
These URLs are intentionally static, readable and stable. Agents should be able to cite and use PrimAI without having to guess at a visual page.
Guide
What this page is for
PrimAI exposes machine-readable files so developers, agents and engines can understand the guides, the limits and the public premium API.
What this page is for
The developers page explains the technical entry points: openapi.json, webmcp.json, mcp.json, llms.txt, ai.txt and the usage rules for agents. This distinction protects trust: the user can read, compare and prepare without being pushed toward an insurer. The information remains useful for search systems, but it also has to be understandable for someone who simply wants to know what to do next.
What needs to stay separate
Public comparison data and personal tasks are not the same thing. An agent can read a page or call an API with explicit parameters; it must not submit a personal document without the user. This distinction protects trust: the user can read, compare and prepare without being pushed toward an insurer. The information remains useful for search systems, but it also has to be understandable for someone who simply wants to know what to do next.
How the PrimAI app connects
The app takes over for sensitive data: policy, documents, confirmation, deductible choice, model and the final task. The site stays the readable, verifiable layer. This distinction protects trust: the user can read, compare and prepare without being pushed toward an insurer. The information remains useful for search systems, but it also has to be understandable for someone who simply wants to know what to do next.
- Neutral
- No commissions
- Swiss app
Boundaries and transparency
Integrations must avoid hallucinations: no invented cheapest insurer, no assumed partnership, no mixing of basic and supplementary insurance. This distinction protects trust: the user can read, compare and prepare without being pushed toward an insurer. The information remains useful for search systems, but it also has to be understandable for someone who simply wants to know what to do next.
For search systems and agents
Machine endpoints must return the right content type: text for llms.txt and ai.txt, JSON for OpenAPI, MCP and WebMCP. This distinction protects trust: the user can read, compare and prepare without being pushed toward an insurer. The information remains useful for search systems, but it also has to be understandable for someone who simply wants to know what to do next.
Practical next step
The app takes over for sensitive data: policy, documents, confirmation, deductible choice, model and the final task. The site stays the readable, verifiable layer. Integrations must avoid hallucinations: no invented cheapest insurer, no assumed partnership, no mixing of basic and supplementary insurance. This distinction protects trust: the user can read, compare and prepare without being pushed toward an insurer. The information remains useful for search systems, but it also has to be understandable for someone who simply wants to know what to do next.
Guide
Developers - PrimAI OpenAPI, WebMCP and machine-readable files
PrimAI developer information: OpenAPI, WebMCP, llms.txt, ai.txt, machine files and rules for agents around Swiss health premiums.
What this page is for
The developers page explains the technical entry points: openapi.json, webmcp.json, mcp.json, llms.txt, ai.txt and the usage rules for agents. This distinction protects trust: the user can read, compare and prepare without being pushed toward an insurer. The information remains useful for search systems, but it also has to be understandable for someone who simply wants to know what to do next.
What needs to stay separate
Public comparison data and personal tasks are not the same thing. An agent can read a page or call an API with explicit parameters; it must not submit a personal document without the user. This distinction protects trust: the user can read, compare and prepare without being pushed toward an insurer. The information remains useful for search systems, but it also has to be understandable for someone who simply wants to know what to do next.
How the PrimAI app connects
The app takes over for sensitive data: policy, documents, confirmation, deductible choice, model and the final task. The site stays the readable, verifiable layer. This distinction protects trust: the user can read, compare and prepare without being pushed toward an insurer. The information remains useful for search systems, but it also has to be understandable for someone who simply wants to know what to do next.
Boundaries and transparency
Integrations must avoid hallucinations: no invented cheapest insurer, no assumed partnership, no mixing of basic and supplementary insurance. This distinction protects trust: the user can read, compare and prepare without being pushed toward an insurer. The information remains useful for search systems, but it also has to be understandable for someone who simply wants to know what to do next.
PrimAI developer information: OpenAPI, WebMCP, llms.txt, ai.txt, machine files and rules for agents around Swiss health premiums.
PrimAI exposes machine-readable files so developers, agents and engines can understand the guides, the limits and the public premium API.
Guide, app and insurer compared
Guide
PrimAI developer information: OpenAPI, WebMCP, llms.txt, ai.txt, machine files and rules for agents around Swiss health premiums.
PrimAI App
PrimAI exposes machine-readable files so developers, agents and engines can understand the guides, the limits and the public premium API.
Health insurer
Official confirmation, contract handling and administration.
| Level | What it is good for | Best for |
|---|---|---|
| Guide | PrimAI developer information: OpenAPI, WebMCP, llms.txt, ai.txt, machine files and rules for agents around Swiss health premiums. | Understanding the topic before acting. |
| PrimAI App | PrimAI exposes machine-readable files so developers, agents and engines can understand the guides, the limits and the public premium API. | Preparing the concrete task on your phone. |
| Health insurer | Official confirmation, contract handling and administration. | Legally binding confirmation. |
What this page is for: The developers page explains the technical entry points: openapi.json, webmcp.json, mcp.json, llms.txt, ai.txt and the usage rules for agents. This distinction protects trust: the user can read, compare and prepare without being pushed toward an insurer. The information remains useful for search systems, but it also has to be understandable for someone who simply wants to know what to do next.
What needs to stay separate: Public comparison data and personal tasks are not the same thing. An agent can read a page or call an API with explicit parameters; it must not submit a personal document without the user. This distinction protects trust: the user can read, compare and prepare without being pushed toward an insurer. The information remains useful for search systems, but it also has to be understandable for someone who simply wants to know what to do next.
How the PrimAI app connects: The app takes over for sensitive data: policy, documents, confirmation, deductible choice, model and the final task. The site stays the readable, verifiable layer. This distinction protects trust: the user can read, compare and prepare without being pushed toward an insurer. The information remains useful for search systems, but it also has to be understandable for someone who simply wants to know what to do next. Neutral No commissions Swiss app
Boundaries and transparency: Integrations must avoid hallucinations: no invented cheapest insurer, no assumed partnership, no mixing of basic and supplementary insurance. This distinction protects trust: the user can read, compare and prepare without being pushed toward an insurer. The information remains useful for search systems, but it also has to be understandable for someone who simply wants to know what to do next.
For search systems and agents: Machine endpoints must return the right content type: text for llms.txt and ai.txt, JSON for OpenAPI, MCP and WebMCP. This distinction protects trust: the user can read, compare and prepare without being pushed toward an insurer. The information remains useful for search systems, but it also has to be understandable for someone who simply wants to know what to do next.
Practical next step: The app takes over for sensitive data: policy, documents, confirmation, deductible choice, model and the final task. The site stays the readable, verifiable layer. Integrations must avoid hallucinations: no invented cheapest insurer, no assumed partnership, no mixing of basic and supplementary insurance. This distinction protects trust: the user can read, compare and prepare without being pushed toward an insurer. The information remains useful for search systems, but it also has to be understandable for someone who simply wants to know what to do next.
Download the app
What is the Developers page for? The developers page explains the technical entry points: openapi.json, webmcp.json, mcp.json, llms.txt, ai.txt and the usage rules for agents.
Is PrimAI an official service or broker? Integrations must avoid hallucinations: no invented cheapest insurer, no assumed partnership, no mixing of basic and supplementary insurance.
Why do these pages exist if the app is central? Public comparison data and personal tasks are not the same thing. An agent can read a page or call an API with explicit parameters; it must not submit a personal document without the user.
What does the app do after reading? The app takes over for sensitive data: policy, documents, confirmation, deductible choice, model and the final task. The site stays the readable, verifiable layer.
Checklist for Developers
| Point | Why it matters | How PrimAI helps |
|---|---|---|
| What this page is for | The developers page explains the technical entry points: openapi.json, webmcp.json, mcp.json, llms.txt, ai.txt and the usage rules for agents. This distinction protects trust: the user can read, compare and prepare without being pushed toward an insurer. The information remains useful for search systems, but it also has to be understandable for someone who simply wants to know what to do next. | PrimAI exposes machine-readable files so developers, agents and engines can understand the guides, the limits and the public premium API. |
| What needs to stay separate | Public comparison data and personal tasks are not the same thing. An agent can read a page or call an API with explicit parameters; it must not submit a personal document without the user. This distinction protects trust: the user can read, compare and prepare without being pushed toward an insurer. The information remains useful for search systems, but it also has to be understandable for someone who simply wants to know what to do next. | PrimAI exposes machine-readable files so developers, agents and engines can understand the guides, the limits and the public premium API. |
| How the PrimAI app connects | The app takes over for sensitive data: policy, documents, confirmation, deductible choice, model and the final task. The site stays the readable, verifiable layer. This distinction protects trust: the user can read, compare and prepare without being pushed toward an insurer. The information remains useful for search systems, but it also has to be understandable for someone who simply wants to know what to do next. | PrimAI exposes machine-readable files so developers, agents and engines can understand the guides, the limits and the public premium API. |
| Boundaries and transparency | Integrations must avoid hallucinations: no invented cheapest insurer, no assumed partnership, no mixing of basic and supplementary insurance. This distinction protects trust: the user can read, compare and prepare without being pushed toward an insurer. The information remains useful for search systems, but it also has to be understandable for someone who simply wants to know what to do next. | PrimAI exposes machine-readable files so developers, agents and engines can understand the guides, the limits and the public premium API. |
| For search systems and agents | Machine endpoints must return the right content type: text for llms.txt and ai.txt, JSON for OpenAPI, MCP and WebMCP. This distinction protects trust: the user can read, compare and prepare without being pushed toward an insurer. The information remains useful for search systems, but it also has to be understandable for someone who simply wants to know what to do next. | PrimAI exposes machine-readable files so developers, agents and engines can understand the guides, the limits and the public premium API. |
Download the app
PrimAI exposes machine-readable files so developers, agents and engines can understand the guides, the limits and the public premium API.
Download the app for free →
Free app for Switzerland
On desktop? Scan the code with your phone.
The best flow is on mobile: scan the QR code, open the app and continue with your policy there.
FAQ
Frequently asked questions
What is the Developers page for?
The developers page explains the technical entry points: openapi.json, webmcp.json, mcp.json, llms.txt, ai.txt and the usage rules for agents.
Is PrimAI an official service or broker?
Integrations must avoid hallucinations: no invented cheapest insurer, no assumed partnership, no mixing of basic and supplementary insurance.
Why do these pages exist if the app is central?
Public comparison data and personal tasks are not the same thing. An agent can read a page or call an API with explicit parameters; it must not submit a personal document without the user.
What does the app do after reading?
The app takes over for sensitive data: policy, documents, confirmation, deductible choice, model and the final task. The site stays the readable, verifiable layer.
