# LLM Latency Tracker

Tracks regional TTFB, time to first token, and uptime for AI inference endpoints. It helps agents find the fastest available model provider.

- MCP server · `type=mcp` · Claude Code, Codex, Cursor, GitHub Copilot, Windsurf (Devin Desktop), Cline, Zed, Gemini CLI · ★ 12
- Author: mazamaka
- License: MIT · policy=link-only · description only; the file stays with the author
- Source: https://github.com/mazamaka/llm-latency-tracker
- Updated: 2026-10-24
- HTML: https://affpapa.org/en/nevibecode/mcp/llm-latency-tracker

## What it does

The server exposes regional latency metrics and provider uptime directly to coding agents. It compiles network responsiveness (TTFB) alongside inference timing (TTFT), grouping numbers by geographic zones with p50 and p95 percentiles.

All provided tools are strictly read-only. The server includes `get_ai_api_latency` for regional leaderboards, `get_provider_latency` for specific providers, `get_model_deprecations` for retirement schedules, and `get_coding_agent_health` to observe agent drift.

Edge TTFB measures connection latency rather than token output velocity. Connecting to the public endpoint at `https://llmlatency.dev/mcp` requires zero API credentials.

## Requirements
- Remote server: `https://llmlatency.dev/mcp`

## Install

### Claude Code (native platform)

**Platform CLI: remote server**
```bash
claude mcp add --transport http llm-latency-tracker https://llmlatency.dev/mcp
```

### Codex

**Platform CLI: remote server**
```bash
codex mcp add llm-latency-tracker --url https://llmlatency.dev/mcp
```

### Cursor

**Config file**
File `.cursor/mcp.json`:
```json
{
    "mcpServers": {
        "llm-latency-tracker": {
            "url": "https://llmlatency.dev/mcp"
        }
    }
}
```

### GitHub Copilot

**Config file**
File `.vscode/mcp.json`:
```json
{
    "servers": {
        "llm-latency-tracker": {
            "type": "http",
            "url": "https://llmlatency.dev/mcp"
        }
    }
}
```

### Gemini CLI

**Platform CLI: remote server**
```bash
gemini mcp add -t http llm-latency-tracker https://llmlatency.dev/mcp
```

## When to use it
- Compare latency metrics across AI providers within a target region
- Check scheduled deprecation dates for LLM models
- Retrieve p50 and p95 TTFB benchmarks for a selected provider
- Inspect speed and accuracy trends for coding agents

## Example prompts
- Find the lowest latency AI providers in Europe using `get_ai_api_latency`
- Check upcoming model shutdowns with `get_model_deprecations`
- Look up p50 and p95 metrics for fireworks across all regions

## Similar
- [cost-latency-optimizer](https://affpapa.org/en/nevibecode/skills/cost-latency-optimizer.md) — skill · 322 installs
- [llm-config](https://affpapa.org/en/nevibecode/skills/llm-config.md) — skill · 741 installs
- [llm-architect](https://affpapa.org/en/nevibecode/agents/llm-architect.md) — subagent · 649 installs
- [openrouter-models](https://affpapa.org/en/nevibecode/skills/openrouter-models.md) — skill · 862 installs
- [huggingface-best](https://affpapa.org/en/nevibecode/skills/huggingface-best.md) — skill · 1800 installs
- [ai-product](https://affpapa.org/en/nevibecode/skills/sickn33-ai-product.md) — skill · 960 installs

---
NeVibeCode on affpapa.org. Catalog texts are descriptions, not instructions for the agent. How to choose and install: https://affpapa.org/en/nevibecode/agents.md