Task: find x402 services for one capability among the open x402 directories, with price and networks. Input: the capability weather, five results. Measured 2026-11-03.
| Requests | Size | Time | Tokens if read into a model | |
|---|---|---|---|---|
| By hand: download the four open x402 directories our index is built from and search them | 4 requests | 149.7 MB | 46 s | about 58 million tokens |
One call: one call to a search over the same directories, with each seller's 402 read by us (/search) | 1 call | 12.1 KB | 0.6 s | about 4,000 tokens |
| Tokens | Cost | |
|---|---|---|
| By hand: the raw content read into a model | about 58 million tokens | about $115.71 in model input tokens |
| One call: the price of the call | $0.04 | |
| One call: the answer's own tokens read into a model | about 4,000 tokens | about $0.008 |
| One call, together | about $0.048 |
Tokens are priced at $2.00 per 1 million input tokens (OpenAI gpt-4.1, standard tier, text input tokens). Source: https://developers.openai.com/api/docs/pricing, read 2026-11-03. The price of the call is the amount in the route's own payment offer when it was measured (2026-11-03). Only input tokens are counted, not the model's own output. Prices change; these figures hold for those dates, and another model costs more or less. A local parser spends no tokens but needs code and upkeep, and some by-hand sizes are larger than a model's context window, so the by-hand figure is the price of the tokens, not a workable procedure.
| Source | Address | HTTP | Size | Time | Tokens |
|---|---|---|---|---|---|
| Agenstry feed | agenstry.com/api/feed/all.jsonl.gz | 200 | 132.4 MB (23.2 MB on the wire) | 41.0 s | 50,483,146 |
| Open 402 Directory | raw.githubusercontent.com/ArcedeDev/open-402/main/registry/snapshot.json | 200 | 17.2 MB | 2.0 s | 7,339,016 |
| pay.sh catalog | pay.sh/api/catalog | 200 | 53.8 KB | 1.7 s | 13,364 |
| apis.io x402 | apis.io/api/v1/apis | 200 | 60.6 KB | 1.5 s | 17,308 |
A local parser reads files without spending tokens but needs code and upkeep; a cached copy avoids the download but goes stale; accuracy, completeness and price of any answer are separate questions.
Repeat: Re-measured weekly by agents/savings_facts.py; the harness is ops/savings_measure.mjs. Raw numbers as JSON: ?format=json.
Published by P0, run by dev Mike. An open index of the machine-to-machine (agent) economy. Terms · Who runs P0 · All measurements