Concepts

How AgentEarth Credits Work: Tool Call Costs Explained

One credit balance settles every call across {{TOOL_COUNT}} tools, so you skip per-API billing. Learn the credit model, how a single call is priced, how caching saves money, and the free credits new users get.

How AgentEarth Credits Work: Tool Call Costs Explained

Tool names in this guide are for reference only and may have changed or been renamed. Please use the latest tools in the tool marketplace when implementing.

Key takeaways

  • Understand how AgentEarth settles every tool call with one credit balance
  • Know what drives the credit cost of a single call and what makes it cheaper or pricier
  • Use caching, precise prompts, and free credits to keep costs down

Tools used

Why "how much does an agent tool call cost" is so hard to estimate

When you wire an AI agent to real capabilities, cost ends up scattered everywhere: a stock-quote API bills per request, image generation charges per image, web scraping charges per page — and every vendor has its own plans, tiers, and minimums. You juggle a pile of invoices, and you still can't tell up front what a given task will cost to run.

AgentEarth collapses all of that into a single unit: credits. One balance settles every call across all 2,400+ tools, so there's no separate account, payment, or reconciliation per API. This guide explains how the credit model works, roughly how a single call is priced, and how to keep spending down.

The credit model: one wallet, every tool

Credits are the platform's unified unit of account. Your account holds a credit balance, and every tool call draws from it. The payoff is that cost becomes predictable and comparable:

  • Whether you call a financial-data tool (such as ae_twelvedata_get_quote), a web search-and-scrape tool (such as ae_serper_search), or an image-generation tool, it all draws down the same credits.
  • You watch one balance instead of a dozen vendor quotas and billing cycles.
  • Switching tools or categories needs no new contract — if your balance covers it, you can call it.

Roughly how a single call is priced

A call isn't a flat fee. What it costs in credits tracks how much work the call actually does. The main dimensions that drive the cost of a single call:

Dimension Effect on cost Intuition
Tool category Light lookups are cheap; heavy compute/generation costs more A quote lookup vs. generating a video
Volume processed More data returned/processed means higher cost Scraping 1 page vs. 50 pages
Upstream resource Heavier underlying paid capability uses more A simple search vs. a large model generation
Call outcome Failed/empty results generally aren't billed as success An upstream error shouldn't cost you

In short: light data lookups are usually cheap, while generation (images, video, long scrapes) is relatively pricier. The exact cost per tool is shown on each tool's detail page and in the on-site pricing notes, so you can check before you run.

How to keep costs down

The same task can cost very differently depending on how you call it. A few practical ways to spend less:

  1. Lean on caching. For identical or near-identical requests, a cache hit generally avoids re-engaging the upstream resource, so repeat queries are much cheaper. Reuse stable, unchanging lookups (reference data, historical values) wherever you can.
  2. Write precise prompts. Rather than letting the agent trial-and-error, scope it in one line: use ae to get AAPL's closing price for the last trading day costs less than "check how Apple's been doing lately," which can trigger several broad calls.
  3. Control the volume. Ask only for the data you need. Capping pages and fields on scrape tasks, and specifying size and count on generation tasks, directly cuts consumption.
  4. Validate on a small sample first. Prove the flow with a neutral sample like example.com, then scale up, so you don't burn credits on an unpolished pipeline.

Free credits for new users

Signing up grants a starting pool of credits — enough to run your common tools end to end and build an intuition for cost before you commit. You don't have to pay to try: validate your workflow on the free credits first, then top up based on what you actually need.

To check usage anytime, view your credit balance and call history in your account to see where spend is going, broken down by tool and time.

How to get started

  1. Sign up — new users get free credits.
  2. Connect in-app (WorkBuddy / Trae), or paste a one-line MCP config into Claude Code / Cursor / VSCode.
  3. Give instructions in natural language, like use ae to check AAPL's live stock price — the agent picks the tool and runs it, drawing credits per actual call.

FAQ

Do new users get free credits? Yes. Signing up grants a starting pool of credits, so you can try common tools without paying first.

Does caching save money? Yes. Identical or near-identical requests that hit the cache generally no longer engage the paid upstream resource, so repeat queries are noticeably cheaper. Reuse results for stable, unchanging data where you can.

How do I check my credit usage? Your account shows your credit balance and call history, broken down by tool and time, so you can see consumption and tune how you call.

Browse 2,400+ tools · Access & pricing notes