Throughput

v2.1.0可观测与分析

opencode-throughput

Real-time LLM performance monitoring plugin for OpenCode. Tracks TTFT, TPS, latency, token usage and cost per model with a TUI sidebar display.

维护活跃

#opencode #opencode-plugin #llm #throughput #latency #ttft

GitHub 仓库 npm 页面官方主页

GitHub 星标

近 30 天 +1

月装机量

658

近 7 天 186

综合评分SCORE

44.3

生态多维模型

快速安装与配置

opencode.json

写入当前项目的 opencode.json，只对这个仓库生效。

opencode.json

{
  "$schema": "https://opencode.ai/config.json",
  "plugin": ["opencode-throughput@2.1.0"]
}

opencode 启动时会通过内嵌运行时自动加载 npm 依赖并缓存至本地目录，无需手动在全局环境执行安装。

README.md

查看原始文件→

Real-time LLM performance monitoring plugin for OpenCode. Tracks latency, throughput, token usage, and cost per model with toast notifications and JSONL logging.

Features

TTFT (Time To First Token) — measures how fast the model starts generating
TPS (Tokens Per Second) — generation throughput during streaming
Total Latency — end-to-end request time
Token Usage — input, output, reasoning, cache read/write
Cost Tracking — per-request and cumulative cost
Toast Notifications — real-time performance summary after each LLM response
JSONL Logging — persistent logs at ~/.opencode/throughput.jsonl
Benchmark Tool — query historical stats on demand from the AI agent

Installation

Add to your opencode.json:

{
  "$schema": "https://opencode.ai/config.json",
  "plugin": ["opencode-throughput"]
}

OpenCode will automatically install the plugin from npm at startup.

Usage

Once installed, the plugin runs automatically:

After each LLM response, a toast notification appears with metrics:
```
claude-sonnet-4 | 2.1s TTFT | 58.3 tok/s | 4.2s | ↑1.2k ↓892
```
Logs are written to ~/.opencode/throughput.jsonl (one JSON object per line)
Query benchmarks — ask the AI agent to use the benchmark tool:

"Show me the benchmark stats for claude" "Compare model performance" "Show the last 10 requests"

Benchmark Tool Args

Arg	Type	Description
`model`	string (optional)	Filter by model name (partial match)
`last`	number (optional)	Only show the last N entries

Log Format

Each line in ~/.opencode/throughput.jsonl:

{
  "ts": "2026-03-21T10:00:00.000Z",
  "model": "anthropic/claude-sonnet-4",
  "providerID": "anthropic",
  "modelID": "claude-sonnet-4",
  "sessionID": "abc123",
  "messageID": "def456",
  "ttft_ms": 2100,
  "tps": 58.3,
  "latency_ms": 4200,
  "inputTokens": 1200,
  "outputTokens": 892,
  "reasoningTokens": 0,
  "cacheReadTokens": 500,
  "cacheWriteTokens": 200,
  "cost": 0.0234,
  "finish": "stop"
}

How It Works

The plugin hooks into OpenCode's event system:

message.updated — captures AssistantMessage.time.completed, tokens, and cost
message.part.updated — captures the first TextPart.time.start for TTFT calculation

Development

Local development

{
  "plugin": ["file:///absolute/path/to/opencode-throughput/src/index.ts"]
}

Build

bun install
bun run build

Publish

npm publish --access public

License

MIT