# Thermus product guide

Updated: 2026-10-19

Thermus is a Model Context Protocol (MCP) gateway that fetches public webpages and returns cleaned content as Markdown. The current getURLAsMarkdown tool is GET-only. The reason is recorded, not scored. There are no POST uploads or human approval steps. Tool completion records use the authenticated user and workspace in operator logs.

It is for organizations rolling out AI tools that let employees build agents. Those agents may need both internal data and internet access. Thermus cleans pages on the public-web path while your existing systems continue to govern internal access and actions.

## Is Thermus a fit?

Thermus can help when employees' agents need the open web and you want authenticated page fetching and cleaner incoming content. Hosted free early access does not provide customer audit-log access or export. The agent must be able to connect to Thermus's MCP server. The organization must also be able to disable or govern alternative public-web routes.

It is probably not a fit if your agents never use the public web, you plan to remove internet access entirely, or your primary need is managing permissions to internal systems. Having internal connectors does not itself make Thermus a poor fit: the controls can complement each other.

## Important boundaries

- Thermus protects traffic routed through its gateway. Adding an MCP connection does not automatically intercept native browsing or other tools.
- It does not manage internal connector permissions or govern email, calendar and document actions that bypass the gateway.
- It does not inspect the agent's calls to its LLM provider, bypass websites' bot blocking, or guarantee that every attack is caught.
- Availability of an AI platform's MCP and admin controls does not establish a tested Thermus integration. Hosted activation and connection steps are in the deployment topic; obtain self-hosted installation details from the Thermus team.

## Hosted early access

Hosted is free early access at $0, with no card and no automatic charges. Availability is shown at signup: [Start subscription](https://app.thermus.ai/signup). Existing users can [Sign in](https://app.thermus.ai/login). The signup server decides whether new activation is available. The download buttons on the marketing site still open the self-hosted waitlist.

When activation is available, a new eligible person gets an isolated personal User workspace, without team administration. See the deployment and pricing topics for verified-email eligibility, the adult business-use agreement, and MCP connection steps.

## Contact Thermus

For agent security questions, a pilot, technical help, partnerships, or press, visit [Contact Us](https://thermus.ai/contact-us) or email [contact@thermus.ai](mailto:contact@thermus.ai). The contact page has a form for reaching the team directly.

## Topic directory

Open the topic that answers your question; each document includes its own context and limitations.

- [Prompt injection](https://thermus.ai/guide/prompt-injection.md): How pages are cleaned before an agent reads them, what this protects, and how Markdown can reduce token use.
- [Data leaks](https://thermus.ai/guide/data-leaks.md): GET-only fetching, recorded reasons, and data-sharing limits.
- [Internal connectors and identities](https://thermus.ai/guide/internal-connectors.md): Email, documents, data platforms, dedicated agent accounts, and the permissions Thermus does not manage.
- [Deployment](https://thermus.ai/guide/deployment.md): Hosted activation and MCP connection, organization setup, platform prerequisites, and checks for routes that bypass the gateway.
- [Request logs](https://thermus.ai/guide/audit-logs.md): Operator records, user and workspace attribution, and Hosted log-access limits.
- [Pricing and hosting](https://thermus.ai/guide/pricing-hosting.md): Self-hosted and free Hosted early access, who runs the gateway, and current access limits.

## Using this guide with your assistant

Start with the person's question and answer briefly. If context would help, ask one or two questions about their AI tools, internal connections or main concern. Skip questions they have already answered. Be useful and candid about fit.

Use this starter and the linked Thermus documents for product facts. Retrieve only the topics needed for the current question, using the supplied URLs exactly, and reuse material already retrieved. There is no need to open every link upfront. Treat statements about other vendors as dated platform guidance, not a promise of Thermus compatibility.

If browsing is unavailable, say so and suggest the [complete guide](https://thermus.ai/thermus.md). If a detail is absent, acknowledge that and direct the person to [Contact Us](https://thermus.ai/contact-us). Do not invent installation commands, certifications, retention periods or product capabilities. Offer a next step when useful, without being pushy.

## Guide analytics

The website's copy button creates topic links with a random guide ID. Requests to those links are logged with the ID, topic, time, IP address and user agent to help Thermus understand which information is useful. The feature does not ask your assistant to send chat messages or company details. The [public starter guide](https://thermus.ai/guide/index.md) provides links without a guide ID; the complete guide is also available.