Skip to content

About

AI Server Manager: public installer and user guides for self-hosted AI, LLMs, GPU inference, images, video and creative plugins. English first; Russian documentation included.

Topics

Resources

Stars

2 stars

Watchers

1 watching

Forks

Latest commit

 

History

10 Commits

Folders and files

NameName
Last commit message
Last commit date
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

AI Server Manager

Русская версия · User documentation · Website

Run and manage AI on your own Linux server. AI Server Manager brings model downloads, GPU assignment, local inference, chat, media generation and automation into one web interface.

This is the public installation and user-documentation repository, not the application source repository. It contains no application source code, model weights, license keys or service credentials. English is the primary language; a separate Russian guide is available.

What you can do

  • Find and launch models: a categorized catalog, downloads, hardware-aware recommendations, saved launch settings and presets.
  • Run local LLMs: vLLM and llama.cpp, with additional experimental integrations such as 1Cat-vLLM, Hipfire and Strata in compatible image variants.
  • Use NVIDIA or AMD GPUs: automatic or manual GPU assignment, engine-specific multi-GPU placement and supported CPU/RAM offload modes.
  • Cluster multiple servers over P2P: connect trusted worker nodes, monitor their GPUs and distribute model inference across servers using supported llama.cpp RPC or vLLM/Ray multi-node configurations. See multi-server operation.
  • Chat and work with an AI agent: model conversations, supported tool calls, code previews and project workflows.
  • Generate images and video: supported FLUX Klein, Ideogram, Wan, MiniMax H3 and LTX pipelines, with resolution presets, progress and saved results.
  • Work with audio and 3D: supported speech recognition, speech synthesis, document/OCR plugins and specialized 3D workflows.
  • Automate: MCP, Telegram, scheduling, knowledge/RAG and other plugins.
  • Operate your server: monitoring, user permissions, API keys and a separate updater with recovery access.

Features depend on the installed image, model format, hardware and license. Experimental presets are not a promise of successful inference on every GPU.

Interface preview

English-language panel screens with mock data. GPU load, memory, models and server details are illustrative, not benchmark results. No production accounts, credentials or server addresses are shown.

Sign in

AI Server Manager sign-in screen in English

Dashboard — blue theme

English dashboard in the blue theme with demo resource charts and GPU monitoring

Dashboard — dark theme

English dashboard in the dark theme with two CMP 170HX GPUs and a modified 20 GB CMP 50HX

Plugin catalog

The catalog includes these 19 implemented plugins. Availability and pricing depend on your subscription, installed build and each plugin's requirements; check AI Store in your panel for current access. Model weights are separate.

Plugin Purpose
MCP Server Connect external AI agents to supported server and model-management tools
Telegram Sentinel Remote server monitoring and control through a Telegram bot
Realtime Voice Agent Interactive voice conversations with speech recognition and synthesis
Cron Task Scheduler Scheduled tasks, model wake-ups, reports and maintenance
SecOps Pentest Agent Defensive security assistant with persistent knowledge and controlled integrations
Docker Manager Container status, logs, lifecycle actions and configuration
Level 0 Support AI First-line support workflows and knowledge-assisted answers
AI Chat Agent AI assistant with knowledge/RAG, web search and supported server tools
Server Messenger Server messaging, rooms, group conversations and direct messages
TG Commenter Pro AI-assisted Telegram commenting, accounts and schedules
AiSharkOCR Document recognition and structured parsing in Lite, Vision and Hybrid modes
Transcriber Agent Audio transcription, speaker diarization and LLM summaries
Object Detector Object detection and segmentation for images and cameras; custom training
LiveKit SIP SIP telephony and voice-agent call integration
AI Pearl Mining Mining runtime management, statistics, console and supported NVIDIA GPU tuning
AI Knowledge Shared knowledge bases, RAG, document versions and access permissions
AI Music Clip Creator Music-driven storyboards, scene generation and music-video export
AI Ad Creator Product/person references, advertising scripts and video export
AI Film Creator Short-film projects, scenes, dialogue, music, effects and export

Planned / prototype catalog entries — not available as production plugins: Crypto/Stock Trader Analyst, Media Factory and DevOps Toolkit.

See the plugin guide for setup and creative workflows.

Creative plugins included in Business

These plugins are included in Business and Enterprise without an additional plugin charge:

Plugin Purpose
AI Music Clip Creator Music track → storyboard → H3/LTX scenes → exported music video
AI Ad Creator Person/product photos and a script → advertising video, Reels or YouTube
AI Film Creator Short-film projects up to 30 minutes, scene references, dialogue/TTS, music, sound effects and export

They support saved projects, reusable character assets, queues and creative MCP tools. Long films are assembled from short scenes, not generated in one request. Actual quality, generation time and lip synchronization require verification. See Plugins.

Quick start

Use a Linux host with Docker/Compose and working GPU drivers where applicable. Review the installation guide before proceeding.

Download the public entry point and read it before running:

curl -fL --proto '=https' --proto-redir '=https' \
  https://raw.githubusercontent.com/rjohny55/ai-server-manager/main/install.sh \
  -o install.sh
bash install.sh --help
bash install.sh auto --download-only

--download-only saves the current official installer as a new file for review. It does not install anything. To install after review:

bash install.sh auto

You can select nvidia, amd or cpu instead of auto. Administrative privileges may be required by the official installer. It creates the panel and the separate updater and may install Docker. Do not run it blindly over an existing production deployment.

The entry point retrieves the current installer from https://ai-server-manager.ru/api/install using verified HTTPS. Review and run the saved downloaded file if you need to execute exactly the version you inspected. No license or account credentials are embedded in this repository.

Documentation

Every chapter includes numbered instructions and localized demonstration screenshots. Start with the English guide or Russian guide.

Links and licensing

Project website · Online user guide · Neithos

AI Server Manager is distributed under its product license. A public installer and public documentation do not make the application open source. Third-party engines and model weights have their own licenses; verify their terms before use. See NOTICE.

For support, open a GitHub issue with a sanitized description. Never post passwords, tokens, license strings, cookies, .env files or private documents.

About

AI Server Manager: public installer and user guides for self-hosted AI, LLMs, GPU inference, images, video and creative plugins. English first; Russian documentation included.

Topics

Resources

Stars

2 stars

Watchers

1 watching

Forks

Releases

Packages

Contributors

Languages