Is my agent's context encrypted before leaving my machine?

Yes. The Privatemode proxy runs locally and encrypts all data before it leaves your machine. It also performs remote attestation to verify the integrity of the Privatemode service. Hermes sends its memory notes and the recent conversation with every request, so all of that travels encrypted.

Does Privatemode store or retain data from my agent's sessions?

No. Privatemode never learns from your data. All requests are processed using confidential computing, and prompts and responses are not saved once a request completes. Only operational metadata such as token usage, timestamps, and request status is kept for billing and monitoring. Hermes itself keeps its memory in `~/.hermes/memories/` and its session history in `~/.hermes/state.db` on your own machine.

Do all Hermes features work with Privatemode?

Everything that uses your configured model goes through Privatemode, including chat, skills, and tool use. By default, Hermes also sends its background tasks, such as context compression and summaries, to your main model, so they are covered too. Tools that call other services, such as web search or image generation, use whichever provider you set up for them.

How can I speed up long agent sessions?

Hermes sends the whole conversation with every request and keeps the start of its prompt stable within a session so that a cache can be reused. Start the proxy with `--sharedPromptCache` to benefit from that. Prompt caching is off by default because all clients of a proxy share the cache once it is on, so only enable it on a proxy that you use alone. Response times are otherwise comparable to standard cloud AI services.

Which model should I use, and how far does the free tier go?

Privatemode offers a range of open-weight models inside confidential computing environments. For agent work, GLM-5.3 is a good start: it has a 256k context window and is well suited for long-running tool use. It is currently in preview. See the [model overview](https://privatemode.ai/models) for all models. Agents send a lot of context with every turn, so the free tier is best for trying things out. For daily use, take a look at the [pricing](https://privatemode.ai/pricing).

Does Privatemode control what the agent does on my computer?

No. Privatemode protects the requests Hermes sends to the model, not what the agent does on your computer. By default, Hermes screens dangerous commands and asks you when it is unsure, and it can run its tools in a container. Read the Hermes [security guide](https://hermes-agent.nousresearch.com/docs/user-guide/security) and decide how much freedom to give it.

Integrating Privatemode AI with Hermes Agent

Hermes Agent is an open-source agent from Nous Research that keeps notes about you, writes its own skills, and answers in your terminal or in messaging apps like Telegram, Discord, and Slack. With Privatemode as its model provider, everything it reads and remembers is processed by open-weight models in an environment whose memory stays encrypted.

Introduction

Let your agent remember everything without exposing it

Encryption at rest, in transit, and in use

The challenge

Your agent's memory goes to the model with every request

Hermes Agent builds up a profile of you, keeps notes on your projects, and can run commands on your machine. All of that reaches the model as part of the prompt, again and again. With a conventional provider, your memory notes, files, and terminal output are processed in plaintext on the provider's servers.

The solution

Encrypted before it leaves your machine

Hermes accepts any OpenAI-compatible API as a custom endpoint, and Privatemode provides one. The Privatemode proxy on your machine encrypts every request before it leaves, and the request is processed inside a hardware-isolated environment whose memory stays encrypted. By design, neither the infrastructure provider nor Privatemode can read your agent's context. This is verifiable through remote attestation and open-source code.

Benefits

Why use Privatemode AI with Hermes Agent?

Works as a custom endpoint

End-to-end confidentiality

Learn how

Free tier available

How to get started

How to set up Privatemode in Hermes Agent

Get your API key

If you don't have a Privatemode API key yet, you can generate one for free here.

Run the Privatemode proxy

The proxy verifies the integrity of the Privatemode service using remote attestation. The proxy also encrypts all data before sending and decrypts data it receives.

Install Hermes Agent

Run the installer on Linux, macOS, or WSL2. For Windows PowerShell and other options, see the official docs.

Connect Hermes to the proxy

Run hermes model, choose Custom endpoint, and enter http://localhost:8080/v1 as the base URL, any placeholder as the API key because the proxy handles authentication, and a model name such as glm-5.3. Hermes saves your choice in ~/.hermes/config.yaml, where you can also set it by hand.

Start Hermes

Run hermes for a session in your terminal. To reach the agent from Telegram, Discord, Slack, or another messaging app, run hermes gateway setup. From now on, every model request goes through Privatemode.

FAQ

Frequently asked questions about using Privatemode with Hermes Agent

Minimal. Hermes accepts any OpenAI-compatible API as a custom endpoint. Point it at the Privatemode proxy with hermes model, or set provider: custom and base_url under model in ~/.hermes/config.yaml. Your memory, skills, and gateway configuration stay as they are.