Inspect every single request.
Don't let your AI be a black box. Track full prompts, responses, and provider metadata with a beautiful, developer-first interface.
| Feature | Model | Provider | Tokens | Latency | Cost | User | Logged | View |
|---|---|---|---|---|---|---|---|---|
| customer-support | gpt-4o | openai | 1.7k(1250+450) | 1450ms | $0.008250 | user_2vXk... | Jul 28, 03:27 AM | |
| search-indexer | claude-3-5-sonnet | anthropic | 5.4k(4200+1200) | 2800ms | $0.0210 | system | Jul 28, 02:27 AM | |
| cover-letter-gen | gpt-4o-mini | openai | 1.4k(850+600) | 1200ms | $0.000210 | user_9mLa... | Jul 28, 01:27 AM |
Includes full support for Reasoning Content,Prompt Caching, and Multi-modal responses.
Pricing Accuracy
Official Provider Support
acost tracks real-time pricing from the world's leading AI providers. Get 100% accurate financial visibility for:
OpenAI
Anthropic
OpenRouter
xAI (Grok)
DeepSeek
Qwen
Minimax
AI Profitability Intelligence
Stop guessing your margins. acost gives you the granular data you need to optimize your AI costs and improve unit economics.
Granular Telemetry
Track provider, model, input/output tokens, cost, and latency for every single request in real-time.
Feature Tracking
Connect AI costs directly to your product features. Know which parts of your app are profitable and which aren't.
Cost Analysis
Beautifully visualized cost breakdowns by model, provider, and feature tag. Identify spend anomalies instantly.
Simple Integration
Generate a workspace key and send telemetry via a simple REST API. No complex setup or proxying required.
Non-Blocking
Send telemetry asynchronously from your backend. Your AI responses stay fast, even if the analytics call fails.
Usage Trends
Monitor growth trends and cost spikes over time. Get ahead of your AI bill before it becomes a problem.
Connect in 3 minutes
acost was built to be invisible to your users and painless for your developers.
Get Your API Key
Create a secure workspace key in seconds and add it to your server's environment variables.
Send Telemetry
After your AI response finishes, send the metadata to our ingest endpoint via a simple, non-blocking POST request.
Optimize & Save
Get instant visibility into costs, token usage, and margins. Use AI-driven insights to cut spend immediately.
Common Questions
acost acts as a lightweight observer. The flow is designed to be non-blocking and highly accurate:
Observation
After your AI call finishes, you send the model ID and token counts to our /track endpoint.
Matching
Our engine matches your request against our global database (synced daily from PriceToken or OpenRouter).
Calculation
We apply the following formula to determine the exact USD cost:
Dual-Source Intelligence
- • Standard: Uses official provider rates via PriceToken.ai.
- • OpenRouter: Automatically uses market rates if
provider: "openrouter"is detected.
Yes, acost is provider-agnostic. However, there are pros and cons to using unlisted vendors:
Pro
Total flexibility. Track local models (Ollama), custom wrappers, or internal proxy layers.
Cons
Requires manual calculation. You must send an estimatedCost in your payload for accurate accounting.
We do not store your provider API keys (OpenAI, Anthropic, etc.) on our servers.
For features like the AI Playground where you Bring Your Own Key (BYOK), we apply industry-standard AES-256 encryption. Your keys are only used to facilitate the request and are never persisted in plain text.
Join the whitelist
Join founders who are building profitable AI products with real-time financial visibility. We'll let you know when we're ready for you.