ComputeGauge

Бесплатно

Provides cost intelligence and a reputation scoring system to help AI agents optimize spending through smart model selection and local-to-cloud routing. It enab

автор: ComputeGauge

GitHub

Описание

Provides cost intelligence and a reputation scoring system to help AI agents optimize spending through smart model selection and local-to-cloud routing. It enables real-time cost tracking and rewards agents for making efficient, high-credibility decisions across various LLM providers.

README

The cost intelligence + agent credibility layer for AI agents. Install once, every session is cost-aware and credibility-building.

npm install -g @computegauge/mcp

Why This Exists

AI agents are powerful but expensive. A single Claude Code session can cost $5-50+. A multi-agent workflow can burn through $100 in minutes. And nobody tells the agent — or the human — until the bill arrives.

ComputeGauge MCP makes agents cost-aware AND gives them a credibility score. Any MCP-compatible agent (Claude, Cursor, Windsurf, custom agents) gets:

Cost Intelligence — Know which model is optimal for every task, track spend in real-time
Agent Credibility — Build a reputation score (0-1000) by making smart decisions. Compete on a leaderboard.
Local→Cloud Routing — Detect when local inference isn't good enough, route to cloud, earn credibility for smart routing

The result: agents that spend 40-70% less, build visible credibility, and know when to route to cloud.

Setup — 30 Seconds

Claude Desktop / Claude Code

Add to ~/.config/claude/claude_desktop_config.json:

{
  "mcpServers": {
    "computegauge": {
      "command": "npx",
      "args": ["-y", "@computegauge/mcp"]
    }
  }
}

That's it. Restart Claude. Every conversation now has cost intelligence + credibility tracking.

With Provider API Keys (Enhanced)

{
  "mcpServers": {
    "computegauge": {
      "command": "npx",
      "args": ["-y", "@computegauge/mcp"],
      "env": {
        "ANTHROPIC_API_KEY": "sk-ant-...",
        "OPENAI_API_KEY": "sk-...",
        "COMPUTEGAUGE_BUDGET_TOTAL": "50"
      }
    }
  }
}

With Local Inference (Ollama, vLLM, etc.)

{
  "mcpServers": {
    "computegauge": {
      "command": "npx",
      "args": ["-y", "@computegauge/mcp"],
      "env": {
        "OLLAMA_HOST": "http://localhost:11434",
        "OLLAMA_MODELS": "llama3.3:70b,qwen2.5:7b,deepseek-r1:14b",
        "ANTHROPIC_API_KEY": "sk-ant-...",
        "COMPUTEGAUGE_BUDGET_TOTAL": "50"
      }
    }
  }
}

Cursor

Add to Cursor MCP settings:

{
  "computegauge": {
    "command": "npx",
    "args": ["-y", "@computegauge/mcp"]
  }
}

Tools Reference

Agent-Native Tools (use automatically every session)

Tool	When to Call	What It Does	Credibility
`pick_model`	Before any API request	Returns the optimal model for a task	+8 Routing Intelligence
`log_request`	After any API request	Logs the request cost	+3 Honest Reporting
`session_cost`	Every 5-10 requests	Shows cumulative cost and budget	—
`rate_recommendation`	After completing a task	Rate how well the model performed	+5 Quality Contribution
`model_ratings`	When curious about quality	View model quality leaderboard	—
`improvement_cycle`	At session end	Run continuous improvement engine	+15 Quality Contribution
`integrity_report`	For transparency	View rating acceptance/rejection stats	—

Credibility Tools (the reputation protocol)

Tool	When to Call	What It Does	Credibility
`credibility_profile`	Anytime	View your 0-1000 credibility score, tier, badges	—
`credibility_leaderboard`	To compete	See how you rank vs other agents	—
`route_to_cloud`	After local→cloud routing	Report smart routing decision	+70 Cloud Routing
`assess_routing`	Before choosing local vs cloud	Should this task stay local?	—
`cluster_status`	To check local capabilities	View local endpoints, models, hardware	—

Intelligence Tools (for user questions)

Tool	Description
`get_spend_summary`	User's total AI spend across all providers
`get_budget_status`	Budget utilization and alerts
`get_model_pricing`	Current pricing for any model
`get_cost_comparison`	Compare costs for specific workloads
`suggest_savings`	Actionable cost optimization recommendations
`get_usage_trend`	Spend trends and anomaly detection

Resources

Resource	URI	Description
Config	`computegauge://config`	Current server configuration
Session	`computegauge://session`	Real-time session cost data
Ratings	`computegauge://ratings`	Model quality leaderboard
Credibility	`computegauge://credibility`	Agent credibility profile + leaderboard
Cluster	`computegauge://cluster`	Local inference cluster status
Quickstart	`computegauge://quickstart`	Agent onboarding guide

Prompts

Prompt	Description
`cost_aware_system`	System prompt that makes any agent cost-aware + credibility-building
`daily_cost_report`	Generate a quick daily cost report
`optimize_workflow`	Analyze and optimize a described AI workflow

Agent Credibility System

Every smart decision earns credibility points on a 0-1000 scale:

Category	How to Earn	Points
🧠 Routing Intelligence	Using `pick_model` wisely, avoiding overspec	+8 to +15 per event
💰 Cost Efficiency	Staying under budget, significant savings	+5 to +30 per event
✅ Task Success	Completing tasks successfully	+10 to +25 per event
📊 Honest Reporting	Logging requests, reporting failures honestly	+3 to +10 per event
☁️ Cloud Routing	Smart local→cloud routing via ComputeGauge	+25 to +70 per event
⭐ Quality Contribution	Rating models, running improvement cycles	+5 to +15 per event

Credibility Tiers

Tier	Score	What It Means
⚪ Unrated	0-99	Just getting started
🥉 Bronze	100-299	Learning the ropes
🥈 Silver	300-499	Competent and cost-aware
🥇 Gold	500-699	Skilled optimizer
💎 Platinum	700-849	Elite decision-maker
👑 Diamond	850-1000	Best in class

Earnable Badges

Badge	How to Earn
🌱 First Steps	Complete first session
💰 Cost Optimizer	Save >$10 through smart model selection
📊 Transparency Champion	Log 50+ requests accurately
☁️ Smart Router	Successfully route 10+ tasks to cloud
⭐ Quality Pioneer	Submit 25+ model ratings
🔥 Streak Master	20+ consecutive successful tasks
🥇 Gold Agent	Reach Gold tier (500+ score)
💎 Platinum Agent	Reach Platinum tier (700+ score)
👑 Diamond Agent	Reach Diamond tier (850+ score)
🌐 Hybrid Intelligence	Use both local and cloud models in one session

Local Cluster Integration

ComputeGauge auto-detects local inference endpoints:

Platform	Environment Variable	Default
Ollama	`OLLAMA_HOST`	`http://localhost:11434`
vLLM	`VLLM_HOST`	—
llama.cpp	`LLAMACPP_HOST`	—
TGI	`TGI_HOST`	—
LocalAI	`LOCALAI_HOST`	—
Custom	`LOCAL_LLM_ENDPOINT`	—

Set OLLAMA_MODELS="llama3.3:70b,qwen2.5:7b" (comma-separated) to declare available models.

The Local→Cloud Routing Flow

1. Agent calls assess_routing("code_generation", quality="good")
2. ComputeGauge checks: local llama3.3:70b quality for code_generation = 80/100
3. "Good" quality threshold = 78 → Local model is sufficient!
4. Agent uses local model → saves money → earns credibility for honest assessment

OR:

1. Agent calls assess_routing("complex_reasoning", quality="excellent")
2. ComputeGauge checks: local llama3.3:70b quality for complex_reasoning = 78/100
3. "Excellent" quality threshold = 88 → Quality gap of 10 points → Route to cloud!
4. Agent calls pick_model → gets Claude Sonnet 4 → executes → calls route_to_cloud
5. Agent earns +70 credibility points for smart routing decision

How `pick_model` Works

The decision engine scores every model across three dimensions:

Quality — Per-task-type scores for 14 task types Cost — Real pricing from 8 providers, 20+ models, calculated per-call (log-scale normalization) Speed — Relative inference speed scores

Priority	Quality	Cost	Speed
`cheapest`	20%	70%	10%
`balanced`	45%	35%	20%
`best_quality`	70%	10%	20%
`fastest`	25%	15%	60%

Model Coverage

Provider	Models	Tier Range
Anthropic	Claude Opus 4, Sonnet 4, Sonnet 3.5, Haiku 3.5	Frontier → Budget
OpenAI	o1, GPT-4o, o3-mini, GPT-4o-mini	Frontier → Budget
Google	Gemini 2.0 Pro, 1.5 Pro, 2.0 Flash	Premium → Budget
DeepSeek	Reasoner, Chat	Value → Budget
Groq	Llama 3.3 70B, Llama 3.1 8B	Value → Budget
Together	Llama 3.3 70B Turbo, Qwen 2.5 72B	Value
Mistral	Large, Small	Premium → Budget

Local Models Supported

Model	Quality (general)	Best For
llama3.3:70b	79/100	General tasks, code
qwen2.5:72b	81/100	Code, math, translation
deepseek-r1:70b	80/100	Reasoning, math, code
deepseek-r1:14b	68/100	Budget reasoning
phi3:14b	60/100	Simple tasks
llama3.1:8b	58/100	Classification, simple QA
mistral:7b	58/100	Simple tasks

Environment Variables

Variable	Required	Description
`COMPUTEGAUGE_DASHBOARD_URL`	No	URL of ComputeGauge dashboard
`COMPUTEGAUGE_API_KEY`	No	API key for dashboard access
`COMPUTEGAUGE_BUDGET_TOTAL`	No	Session budget limit in USD
`COMPUTEGAUGE_BUDGET_ANTHROPIC`	No	Per-provider monthly budget
`COMPUTEGAUGE_BUDGET_OPENAI`	No	Per-provider monthly budget
`ANTHROPIC_API_KEY`	No	Enables Anthropic provider detection
`OPENAI_API_KEY`	No	Enables OpenAI provider detection
`GOOGLE_API_KEY`	No	Enables Google provider detection
`OLLAMA_HOST`	No	Ollama inference endpoint
`OLLAMA_MODELS`	No	Comma-separated local model names
`VLLM_HOST`	No	vLLM inference endpoint
`COMPUTEGAUGE_GPU`	No	GPU name for hardware detection
`COMPUTEGAUGE_VRAM_GB`	No	VRAM in GB
`COMPUTEGAUGE_COST_PER_HOUR`	No	Amortized hardware cost/hr

For Agent Developers

If you're building AI agents (via Claude Agent SDK, LangChain, CrewAI, AutoGen, etc.), ComputeGauge MCP is the easiest way to add cost awareness AND agent credibility:

Zero integration effort — Just add the MCP server to your agent's config
No code changes — The agent discovers 18 tools via MCP protocol automatically
Immediate value — pick_model returns recommendations on first call, credibility tracking starts automatically
Session tracking built-in — Full cost visibility per agent run
Credibility system — Your agent earns a visible reputation score that users can see
Local cluster support — Auto-detect and leverage on-prem inference
Budget guardrails — Warnings when approaching limits

Pattern: Cost-Aware + Credibility-Building Agent Loop

1. Agent receives task
2. Agent calls assess_routing(task_type) → local or cloud?
3. Agent calls pick_model(task_type, priority="balanced")
4. Agent uses recommended model for the task
5. Agent calls log_request(provider, model, tokens)
6. Agent calls rate_recommendation(model, rating, success)
7. If cloud-routed: agent calls route_to_cloud(task_type, reason, model)
8. Every 5 requests, agent calls session_cost()
9. If session cost > 80% of budget, switch to priority="cheapest"
10. At session end: check credibility_profile()

This pattern reduces costs by 40-70% while building a credibility score that makes users trust the agent more.

License

Apache-2.0 — Free to use, modify, and distribute.

Как установить

Добавь это в claude_desktop_config.json и перезапусти Claude Desktop.

{
  "mcpServers": {
    "computegauge-mcp": {
      "command": "npx",
      "args": []
    }
  }
}

ComputeGauge

Описание

README

Why This Exists

Setup — 30 Seconds

Claude Desktop / Claude Code

With Provider API Keys (Enhanced)

With Local Inference (Ollama, vLLM, etc.)

Cursor

Tools Reference

Agent-Native Tools (use automatically every session)

Credibility Tools (the reputation protocol)

Intelligence Tools (for user questions)

Resources

Prompts

Agent Credibility System

Credibility Tiers

Earnable Badges

Local Cluster Integration

The Local→Cloud Routing Flow

How `pick_model` Works

Model Coverage

Local Models Supported

Environment Variables

For Agent Developers

Pattern: Cost-Aware + Credibility-Building Agent Loop

License

Links

Как установить

Похожие MCP

Fetch

AWS KB Retrieval

Spring AI MCP Server

llm-analysis-assistant

Command Palette

ComputeGauge

Описание

README

Why This Exists

Setup — 30 Seconds

Claude Desktop / Claude Code

With Provider API Keys (Enhanced)

With Local Inference (Ollama, vLLM, etc.)

Cursor

Tools Reference

Agent-Native Tools (use automatically every session)

Credibility Tools (the reputation protocol)

Intelligence Tools (for user questions)

Resources

Prompts

Agent Credibility System

Credibility Tiers

Earnable Badges

Local Cluster Integration

The Local→Cloud Routing Flow

How pick_model Works

Model Coverage

Local Models Supported

Environment Variables

For Agent Developers

Pattern: Cost-Aware + Credibility-Building Agent Loop

License

Links

Как установить

Похожие MCP

Fetch

AWS KB Retrieval

Spring AI MCP Server

llm-analysis-assistant

How `pick_model` Works