Relagent
Trustworthy AI assistant that runs 100% on your hardware
Relagent is a private and selfhosted AI assistant that runs on your hardware. In addition to this app, you need to run the Relagent engine.
To get started, please have a look at the project's readme at https://gitlab.com/RmMsr/relagent.
The goal of the Relagent project is an accessible solution for agentic AI services with full control over your data. Setting up your own installation should be within reach without considerable technical knowledge or financial investment. The software is designed to be resource-efficient without any hidden costs.
What Relagent brings you:
* 🗨️ Chat interface: Chat with your AI assistant and manage sessions.
* 🌐 Persistent sessions: Open new sessions and continue them on any device. Data stays on your systems only.
* 🎤 Speech Integration: Input and Output on-device with about 40 languages.
* ⚖️ Use any open-weights model: By using the OpenAI-compatible chat completions protocol, you can choose the inference service that works best for you. No external service is needed - but you can use one.
* 🔒️ API Key Authentication: Restrict access to your data. Combine with an HTTPS proxy or a VPN to protect your data while accessing it from anywhere.
* 🧰 Agentic tools: For example, web search. You can decide in which cases you want to be asked for permission based on your selected privacy level.
What you do NOT need:
* Register for a third-party service (unless you want to use a commercial LLM provider)
* Pay for any usage or access (but I appreciate donations)
* Fear that your data ends up in big data silos or model training
What you CAN do:
* Host the Relagent engine yourself (start on your computer, switch to permanent installation when you are ready)
* Run a Large Language Model yourself (needs one additional app, there are several free options with or without a graphical interface to choose from)
* Protect the API with HTTPS, Basic Auth or using a VPN.
Permissions on your device:
* In order to send data between the app and the engine and to download speech models, internet access is needed
* For the voice input (ASR) and output (TTS) features several permissions includiing bluetooth and background processing are required. Relagent won't continuously listen on the microphone unless you explicitly activate it
For more information:
* Find the Relagent source code at https://gitlab.com/RmMsr/relagent.
* Have a look at the related Blog at https://venkado.org/.
* Follow the author @guesser (https://sigmoid.social/@guesser) on the Fediverse.
To get started, please have a look at the project's readme at https://gitlab.com/RmMsr/relagent.
The goal of the Relagent project is an accessible solution for agentic AI services with full control over your data. Setting up your own installation should be within reach without considerable technical knowledge or financial investment. The software is designed to be resource-efficient without any hidden costs.
What Relagent brings you:
* 🗨️ Chat interface: Chat with your AI assistant and manage sessions.
* 🌐 Persistent sessions: Open new sessions and continue them on any device. Data stays on your systems only.
* 🎤 Speech Integration: Input and Output on-device with about 40 languages.
* ⚖️ Use any open-weights model: By using the OpenAI-compatible chat completions protocol, you can choose the inference service that works best for you. No external service is needed - but you can use one.
* 🔒️ API Key Authentication: Restrict access to your data. Combine with an HTTPS proxy or a VPN to protect your data while accessing it from anywhere.
* 🧰 Agentic tools: For example, web search. You can decide in which cases you want to be asked for permission based on your selected privacy level.
What you do NOT need:
* Register for a third-party service (unless you want to use a commercial LLM provider)
* Pay for any usage or access (but I appreciate donations)
* Fear that your data ends up in big data silos or model training
What you CAN do:
* Host the Relagent engine yourself (start on your computer, switch to permanent installation when you are ready)
* Run a Large Language Model yourself (needs one additional app, there are several free options with or without a graphical interface to choose from)
* Protect the API with HTTPS, Basic Auth or using a VPN.
Permissions on your device:
* In order to send data between the app and the engine and to download speech models, internet access is needed
* For the voice input (ASR) and output (TTS) features several permissions includiing bluetooth and background processing are required. Relagent won't continuously listen on the microphone unless you explicitly activate it
For more information:
* Find the Relagent source code at https://gitlab.com/RmMsr/relagent.
* Have a look at the related Blog at https://venkado.org/.
* Follow the author @guesser (https://sigmoid.social/@guesser) on the Fediverse.
版本
雖然在下方可選擇下載 APK 檔案，但要留意這樣的安裝方式將不會收到更新通知，是一種較不安全的下載方法。建議您先安裝 F-Droid 用戶端使用。
下載 F-Droid-
x86_64此版本需要 Android 7.0 或更高的版本。
此套件包由原開發者建置和簽署，並保證與此原始碼 Tarball 保持一致。
-
arm64-v8a此版本需要 Android 7.0 或更高的版本。
此套件包由原開發者建置和簽署，並保證與此原始碼 Tarball 保持一致。
-
armeabi-v7a此版本需要 Android 7.0 或更高的版本。
此套件包由原開發者建置和簽署，並保證與此原始碼 Tarball 保持一致。