Frame Pulse MCP
Monitors Blender rendering processes, detects thermal spikes, and provides tools to throttle or manage Blender workloads to prevent crashes.
Sends rich, color-coded alerts via Discord webhooks for team-wide render farm monitoring.
Monitors Houdini simulations and provides thermal status and emergency throttling capabilities.
Sends private production pager-style notifications via Telegram bot when system thresholds are exceeded.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@Frame Pulse MCP check system health before starting a render"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
in progress
🎬 Frame Pulse MCP
AI-native telemetry for creative workstations.
Give your AI assistant (Claude, Hermes, Cursor) a "nervous system" to monitor and govern heavy creative workloads—Blender renders, Unreal builds, Houdini simulations—preventing thermal crashes and lost work during crunch.

🎯 The Problem Frame Pulse Solves
Digital content creation (DCC) tools are resource-intensive. During overnight renders or UE5 lighting builds, workstations overheat, hang, or crash—destroying hours of work and missing deadlines.
Existing tools (Glances, htop) show data. Frame Pulse lets AI act.
Traditional Monitoring | Frame Pulse with AI |
You watch numbers | Claude asks "Is my system safe?" |
Crash after the fact | Throttle before thermal damage |
Manual Task Manager | Natural language "deprioritize Chrome" |
Silent failure | Discord/Telegram alerts at 2AM |
Related MCP server: System Monitor MCP Server
🚀 Quick Start
Option A: Claude Desktop (Recommended)
Clone and install
git clone [https://github.com/idaakiwumi/frame-pulse-mcp.git](https://github.com/idaakiwumi/frame-pulse-mcp.git)
cd frame-pulse-mcp
pip install -r requirements.txtAdd to Claude Desktop config
Windows (MSIX/Store):
%LOCALAPPDATA%\Packages\Claude_pzs8sxrjxfjjc\LocalCache\Roaming\Claude\claude_desktop_config.json**Other possible paths
\AppData\Roaming\Claude\claude_desktop_config.json\AppData\Local\Claude\claude_desktop_config.jsonStandard:
%APPDATA%\Claude\claude_desktop_config.json
{
"mcpServers": {
"frame-pulse": {
"command": "py",
"args": [
"-3.14",
"-u",
"C:\\Users\\YOUR_PATH\\frame-pulse-mcp\\src\\frame_pulse\\server.py"
]
}
}
}Then ask Claude: "Check my system status" or "Deprioritize my background apps"
Option B: The "Hero Demo" (Streamlit)
I built a high-fidelity Mission Control Dashboard using Streamlit. While the MCP runs in the background, this dashboard serves as the visual command center for the project.
streamlit run streamlit_app/mission_control.py📱 The "Digital Pager" Ecosystem (Telegram & Discord)
As a Product Architect, I designed Frame Pulse to be "Headless." Creative professionals don't want another window to monitor; they want an assistant that pings them when they are away from their desk.
Telegram Bot: Acts as a private production pager. Get a message on your phone if your CPU hits 90°C while you're grabbing coffee.
Discord Webhooks: Rich, color-coded embeds sent to your studio's Discord server for team-wide render farm monitoring.
Setup: Configure your .env file with your TELEGRAM_BOT_TOKEN and TELEGRAM_CHAT_ID.
🛠️ MCP Tools for AI Agents
Tool | Function | Use Case |
| CPU, RAM, and Thermal scan | "Is it safe to start a 6-hour render?" |
| Identifies Blender, Unreal, Maya, etc. | "What's currently taxing the GPU?" |
| Immediate CPU priority reduction | "CRITICAL: Throttle background apps now." |
| Real-time temperature read | "Am I at risk of a thermal shutdown?" |
🤝 Community & Recognition
**Featured on Lobehub — AI tools discovery platform.
🛠️ Built With
MCP — Model Context Protocol for AI tool interoperability
FastMCP — Python SDK for MCP servers
psutil — Cross-platform system monitoring
Streamlit — Rapid Python dashboarding
📸 Demo
[30-second GIF: Claude asking "Start my render?" → Frame Pulse responding "CAUTION: CPU 89%" → User clicking throttle → Success toast]
🤝 Who's Using This
"Frame Pulse caught a thermal spike before our overnight farm render. Saved us 14 hours of redo work."
— Anonymous VFX Supervisor (via DM)
☕ Support the Mission
If this tool saved your render from thermal throttling, prevented a 3AM crash, or helped your AI agent make hardware-aware decisions—consider fueling continued development:
👩🏽💻 Developed by Ida Akiwumi
Lead Product Designer | Creative Technologist | Narrative Strategist
Translating user friction into product opportunities.
Available Tools
10 toolscheck_system_healthA
Check if workstation is safe for heavy creative workloads.
Inputs: no named parameters.
deprioritize_processA
SAME AS THROTTLE — reduces CPU priority of a process.
Use this if "throttle" sounds too aggressive. Same function, clearer language.
Args: target_pid: Process ID to deprioritize. If None, auto-selects highest CPU consumer.
Returns: Confirmation with full explanation of effects.
Inputs: target_pid.
emergency_throttle_aliasD
Alias for throttle_process.
Inputs: target_pid.
emergency_throttle_processA
EMERGENCY: Immediately reduce priority without confirmation.
Use when:
System in CRITICAL thermal state (CPU >90%, Temp >85°C)
Render farm node about to shut down from overheating
No time for explanation
Returns: Action taken, minimal explanation.
Inputs: target_pid.
find_render_processes_aliasC
Alias for scan_creative_apps.
Inputs: app_name.
get_thermal_status_aliasD
Alias for check_system_health.
Inputs: no named parameters.
prioritize_processA
Give a process HIGHER than normal CPU priority.
WARNING: Use sparingly. High priority can make system unresponsive.
Use only when:
A critical render must finish by deadline
System is otherwise idle
You accept risk of temporary unresponsiveness
Args: target_pid: Process ID to boost (REQUIRED)
Returns: Confirmation with warning about side effects.
Inputs: target_pid (required).
restore_priorityA
Restore a throttled process to normal CPU priority.
Use this to 'undo' a throttle/deprioritize if you throttled the wrong process or if your main render finished and you want full performance back.
Args: target_pid: Process ID to restore to normal priority (REQUIRED, no auto-detect for safety)
Returns: Confirmation of priority restoration.
Inputs: target_pid (required).
scan_creative_appsC
Find active creative applications (Blender, Unreal, Maya, etc.).
Inputs: app_name.
throttle_processA
Reduce CPU priority of a process to prevent overheating.
ALIASES: Also responds to "deprioritize" for clarity.
Args: target_pid: Process ID to throttle. If None, auto-throttles highest CPU consumer.
Returns: Confirmation message with explanation of what was done.
Inputs: target_pid.
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 10 tools
Multiple tools are aliases or near-duplicates (e.g., throttle_process, deprioritize_process, emergency_throttle_alias, emergency_throttle_process), making it hard for an agent to distinguish between them without careful reading. Similarly, find_render_processes_alias duplicates scan_creative_apps, and get_thermal_status_alias duplicates check_system_health.
Tool names use a mix of verbs like check, deprioritize, emergency_throttle, find, get, prioritize, restore, scan, throttle with no consistent pattern. The inclusion of '_alias' suffixes alongside action-oriented names adds inconsistency.
The total of 10 tools is within the normal range, but nearly half are redundant aliases, making the set feel bloated. The actual distinct functionality is only 5 operations, suggesting the tool count could be streamlined.
Core operations for CPU priority management are covered (health check, app scanning, throttle/deprioritize, emergency throttle, restore, boost). However, missing generic process listing and ability to set arbitrary priority values are minor gaps that could affect some workflows.
Maintenance
Related MCP Connectors
- DNS-DATA RemoteDeprecatedcom.dns-data
Connect AI assistants to authorized Windows and macOS devices for files, commands, and management.
Windows computers for AI agents. Office, desktop apps, browser tasks, testing and file transfer.
Connect, monitor, and control AI agents — tasks, approvals, schedules, and governance.
Connect AI assistants to your approved computer for files, commands, browser and desktop tasks.
Related MCP Servers
- AlicenseNot gradedqualityCmaintenanceA real-time system diagnostics MCP server that gives AI agents live access to CPU, RAM, disk, network, processes, and hardware health metrics, with zero cloud dependency.7Apache 2.0
- FlicenseNot gradedqualityDmaintenanceGives AI agents real-time access to system metrics, process management, and container orchestration.-
- AlicenseNot gradedqualityDmaintenanceEnables AI agents to monitor system resources and control media, brightness, volume, and other Windows PC functions through natural language commands.1MIT

Yamaru Hardware Probeofficial
AlicenseAqualityDmaintenanceExpert system hardware probe and performance diagnostic engine for AI, Gaming, and High-Performance workflows. Provides deep system insights such as real-time monitoring, thermal diagnostics, and LLM optimization.1157 npm7Apache 2.0