Observability

See everything. Miss nothing.

Full visibility into every LLM request. Logs, cost tracking, latency monitoring, and token analytics across 9 providers.

9
Providers
Real-time
Logs
Cost
Tracking
Fold
Savings
OpenAIAnthropicGoogleCohereMistralGroqOllamaAzureCustom
Features

Complete LLM observability

Request Logging

Every LLM request logged with full context: input, output, tokens, and timing.

Cost Tracking

Per-request cost calculation across all providers. See exactly where money goes.

Latency Monitoring

Track response times by provider, model, and endpoint. Catch regressions early.

Token Analytics

Input vs output token usage over time. Understand your consumption patterns.

Fold Savings

See exactly how much Fold saves you per request, per session, and per month.

Feedback Scoring

Attach quality scores to requests. Correlate prompt changes with response quality.

Integration

One flag to enable everything

Add platform options to your existing Fold setup. Logging starts automatically.

import { fold } from "@fold-run/sdk";

const ctx = fold("coding", {
  platform: {
    apiKey: process.env.FOLD_API_KEY,
    projectId: "my-project",
    // Auto-logs every request with cost, latency, and tokens
    logging: true,
    analytics: true,
  },
});

// Your normal agent loop — Fold logs everything automatically
ctx.system("You are a coding assistant");
ctx.observe(searchResults, "web-search");

const messages = ctx.messages();
// Request logged: provider, model, tokens, cost, latency, fold savings
Dashboard

Your LLM operations at a glance

12,847
Total Requests (24h)
1.2s
Avg Latency
$34.21
Total Cost (24h)
$89.47
Fold Savings (24h)
4.2M
Token Usage
0.02%
Error Rate
Use Cases

Built for every team

Cost Optimization

Identify expensive prompts and optimize them. Track spending by project, team, and model.

Debugging

Replay any request with full context. See exactly what your LLM received and returned.

Compliance

Complete audit trail of every LLM interaction. Filter and export logs by date, user, or project.

Performance Monitoring

Set alerts for latency spikes, cost overruns, or error rate increases. Stay ahead of issues.

See what your LLMs are really doing.

Full observability in minutes.