API Logger

Real-time API traffic logging + multi-provider LLM analysis + TUI dashboard + webhook reporting.

Quick Start

# Build
cargo build --release

# Serve (axum with logging middleware on :3000)
cargo run -- serve
cargo run -- serve --port 4000 --csv ./custom.csv

# Demo client (generates test log entries)
cargo run -- demo-client

# Daemon (HTTP API — integrate from any language)
cargo run -- daemon
cargo run -- daemon --port 9000 --webhook https://hooks.slack.com/...

# TUI dashboard (requires Node.js)
cargo run -- tui
cargo run -- tui --csv ./custom.csv

# Analysis (requires LLM_PROVIDER + API key)
export LLM_PROVIDER=openai
export OPENAI_API_KEY=sk-...
cargo run -- analyze

# Shell completions
cargo run -- completions bash >> ~/.bashrc
cargo run -- completions zsh >> ~/.zshrc

Daemon Endpoints

POST /api/log

curl -X POST http://localhost:8080/api/log \
  -H "Content-Type: application/json" \
  -d '{
    "source": "my-app",
    "method": "GET",
    "endpoint": "/api/users",
    "status_code": 200,
    "latency_ms": 45,
    "request_size": 128,
    "response_size": 4096,
    "error": null
  }'

POST /api/analyze

curl -X POST http://localhost:8080/api/analyze \
  -d '{"provider": "openai", "model": "gpt-4o"}' \
  -H "Content-Type: application/json"

Returns:

{
  "status": "ok",
  "provider": "openai",
  "model": "gpt-4o",
  "analysis": "...",
  "error_rate": 4.2,
  "total_requests": 150,
  "error_count": 6
}

CLI

Usage: api_log_tracker [COMMAND]

Commands:
  serve        Start axum server on :3000 with logging middleware
  demo-client  Make example calls through LoggedClient
  analyze      Run LLM analysis on log entries
  daemon       Start HTTP daemon (default port 8080)
  tui          Launch the TUI dashboard (requires Node.js)
  completions  Generate shell completions
  help         Print this message or the help of the given subcommand(s)

Options:
  -h, --help         Print help
  -V, --version      Print version

Global flags (work on all subcommands):

Analyze-specific:

Daemon-specific:

LLM Providers (13)

Set LLM_PROVIDER + the matching API key:

Optional: LLM_MODEL=gpt-4o to override default model.

Webhook Reporting

cargo run -- daemon --port 8080 --webhook https://hooks.slack.com/...

Triggers when error rate exceeds 5% (configurable via WEBHOOK_THRESHOLD). Posts structured JSON with anomalies, latency hotspots, and recommendations.

Deploy

# Docker
./deploy.sh docker

# Fly.io
./deploy.sh fly

# Railway
./deploy.sh railway

# Render (push to GitHub, connect in dashboard)
./deploy.sh render

# Hugging Face Spaces
./deploy.sh hf

# Local build
./deploy.sh local

Docker Compose starts the daemon on port 8080 with health checks:

docker compose up -d
curl http://localhost:8080/api/health

CSV Log

Default location: logs/api_logs.csv. Override with API_LOGGER_CSV env var.

Cache

LLM analysis results are cached for 60s (keyed by SHA256 of stats). Prevents redundant API calls when multiple clients hit /api/analyze.

TUI Dashboard

# From Rust binary
cargo run -- tui

# Or install the npm package globally
npm install -g api-log-tracker-tui
api-log-tui

# Or from a repo clone
cd my-terminal-ui && npm install && npm run build && node dist/cli.js

Tabs:

• Dashboard — live stats, status bars, sparkline
• Live Logs — scrollable log table with filter
• Analysis — LLM provider selector + run analysis
• Controls — start/stop server, run demo client
• Integration — copy-paste snippets for Python, C, C++, Rust, Go, Erlang + LLM provider env var reference

Use 1-5 to switch tabs, j/k to navigate within tabs, r to run analysis.

Using as a Library

[dependencies]
api_log_tracker = { path = "../api_log_tracker" }

use api_log_tracker::{ApiLogger, LoggedClient};

let logger = ApiLogger::new("logs/api_logs.csv");
let client = LoggedClient::new(logger.clone());

let resp = client.get("https://api.example.com/data").await?;

For axum middleware:

use api_log_tracker::{ApiLogger, middleware::log_requests};

let logger = ApiLogger::new("logs/api_logs.csv");
let app = Router::new()
    .route("/", get(handler))
    .layer(axum::middleware::from_fn_with_state(logger.clone(), log_requests))
    .with_state(logger);

Supported Languages (Integration Snippets)

The TUI Integration tab includes copy-paste snippets for:

All snippets POST to http://localhost:8080/api/log with the same JSON schema.