README
x402station.io Preflight Dataset v0.1
x402station.io Preflight Dataset v0.1 — aggregated probe data from the x402 agentic-commerce network.
This dataset is a weekly public snapshot of endpoint-level aggregate health and risk signals used by Preflight by x402station.io. It is designed for citation, reproducibility, ecosystem analysis, and lightweight agent-routing research without exposing the raw probe-by-probe time-series that powers the paid API.
Canonical service: https://x402station.io
Open signal vocabulary: https://x402station.io/spec#signals
Dataset repository: https://huggingface.co/datasets/x402station/preflight-dataset-v0_1
Files
| File | Format | Contents |
|---|---|---|
data/preflight_dataset_v0_1.csv | CSV | Portable tabular snapshot. |
data/preflight_dataset_v0_1.jsonl | JSONL | Same rows with classification_reasons as an array. |
dataset_manifest.json | JSON | Export metadata, row count, file paths, and boundary policy. |
Schema
| Column | Type | Description |
|---|---|---|
endpoint_url | string or null | Public endpoint URL when no query/userinfo/fragment/sensitive path segment is present. |
endpoint_hostname | string | Hostname retained for every row; used when endpoint_url is redacted. |
endpoint_url_redacted | boolean | true when the full URL was withheld. |
redaction_reason | string or null | Pipe-separated reason: userinfo, query_params, fragment, sensitive_path_segment. |
service_id | string | x402station service identifier from the public catalog surface. |
service_name | string or null | Service display name. |
provider | string or null | Provider/domain grouping used for concentration checks. |
category | string or null | Catalog category when supplied upstream. |
classification | string | Primary class chosen from live, decoy_price_extreme, zombie, dead, dead_7d, mostly_dead, never_paid_zombie, new_provider, slow, price_outlier_high, high_concentration. |
classification_reasons | array or pipe-separated string | All classes that fired for the row. |
first_seen_at | ISO 8601 or null | First time x402station.io saw the endpoint. |
last_probe_at | ISO 8601 or null | Most recent probe included in the aggregate window. |
total_probes_30d | integer | Count of HTTP-naked probes in the trailing 30-day window. |
uptime_30d_pct | number or null | Percent of 30-day probes returning HTTP 200 or 402. |
latency_p50_ms | integer or null | 30-day p50 latency across probes with an HTTP status. |
latency_p95_ms | integer or null | 30-day p95 latency across probes with an HTTP status. |
latency_p99_ms | integer or null | 30-day p99 latency across probes with an HTTP status. |
price_amount_usdc | number or null | Current public catalog price amount, normalized as USDC where supplied. |
price_currency | string or null | Current public catalog currency. |
cdp_l30d_calls | integer or null | Coinbase CDP Bazaar 30-day settlement count when synced. |
cdp_l30d_unique_payers | integer or null | Coinbase CDP Bazaar 30-day unique payer count when synced. |
cdp_last_called_at | ISO 8601 or null | Last CDP-observed settled call timestamp when synced. |
cdp_last_synced_at | ISO 8601 or null | Last CDP quality sync timestamp. |
has_free_trial | boolean or null | Whether CDP enrichment reports a free trial path. |
generated_at | ISO 8601 | Snapshot generation timestamp. |
Probe Methodology
x402station.io probes every indexed x402 endpoint on an approximately 10-minute cadence. A probe is an HTTP naked call: no PAYMENT-SIGNATURE, no paid settlement, and no attempt to consume protected paid content. The probe records aggregate availability evidence such as status class, network failure, and latency.
This public dataset contains endpoint-level aggregates only. It intentionally does not include raw probe-by-probe rows, raw response bodies, payment signatures, wallet identifiers, webhook payloads, or per-request time-series.
Classification Heuristics
The public classes mirror the x402station.io signal vocabulary documented at https://x402station.io/spec#signals.
| Class | Heuristic |
|---|---|
live | No v0.1 dataset class fired. Check total_probes_30d and last_probe_at before treating this as a guarantee. |
decoy_price_extreme | Current price is at least $1,000 USDC. |
zombie | At least 3 probes in the last hour and zero healthy responses. |
dead | At least 3 unhealthy probes in the last 30 minutes. |
dead_7d | At least 20 probes in 7 days and zero healthy responses. |
mostly_dead | At least 20 probes in 7 days and uptime below 50%. |
never_paid_zombie | CDP synced, zero 30-day settled calls, no historical last-called timestamp, endpoint older than 30 days. |
new_provider | Service first seen less than 24 hours before export. |
slow | Last-hour average latency at least 2,000 ms, or 30-day p99 latency at least 5,000 ms. |
price_outlier_high | Current price is at least 10x the provider/domain group median. |
high_concentration | Provider/domain group accounts for at least 5% of active catalog endpoints. |
When multiple classes fire, classification is the highest-priority class and classification_reasons lists the complete set.
Refresh Cadence
The intended public cadence is weekly snapshot refresh. The paid Preflight API remains the fresh path for production agents that need current verdicts, batch checks, and SLA behavior.
License
CC-BY-4.0. Downstream usage should cite x402station.io and this dataset URL.
Citation
Plain text:
x402station.io. x402station.io Preflight Dataset v0.1: aggregated probe data from the x402 agentic-commerce network. Hugging Face Datasets, 2026. https://huggingface.co/datasets/x402station/preflight-dataset-v0_1
BibTeX:
@dataset{x402station_preflight_dataset_v0_1,
title = {x402station.io Preflight Dataset v0.1},
author = {{x402station.io}},
year = {2026},
publisher = {Hugging Face},
url = {https://huggingface.co/datasets/x402station/preflight-dataset-v0_1},
license = {CC-BY-4.0},
note = {Aggregated probe data from the x402 agentic-commerce network}
}
Boundary Policy
Open: endpoint-level aggregate stats, public catalog metadata, classification labels, settlement count/recency aggregates when synced from CDP Bazaar.
Gated: raw probe-by-probe time-series, per-probe response detail, paid API freshness, bulk routing checks, webhook monitoring, and SLA surfaces.
Contributors
3 commits
Similar repos
README
x402station.io Preflight Dataset v0.1
x402station.io Preflight Dataset v0.1 — aggregated probe data from the x402 agentic-commerce network.
This dataset is a weekly public snapshot of endpoint-level aggregate health and risk signals used by Preflight by x402station.io. It is designed for citation, reproducibility, ecosystem analysis, and lightweight agent-routing research without exposing the raw probe-by-probe time-series that powers the paid API.
Canonical service: https://x402station.io
Open signal vocabulary: https://x402station.io/spec#signals
Dataset repository: https://huggingface.co/datasets/x402station/preflight-dataset-v0_1
Files
| File | Format | Contents |
|---|---|---|
data/preflight_dataset_v0_1.csv | CSV | Portable tabular snapshot. |
data/preflight_dataset_v0_1.jsonl | JSONL | Same rows with classification_reasons as an array. |
dataset_manifest.json | JSON | Export metadata, row count, file paths, and boundary policy. |
Schema
| Column | Type | Description |
|---|---|---|
endpoint_url | string or null | Public endpoint URL when no query/userinfo/fragment/sensitive path segment is present. |
endpoint_hostname | string | Hostname retained for every row; used when endpoint_url is redacted. |
endpoint_url_redacted | boolean | true when the full URL was withheld. |
redaction_reason | string or null | Pipe-separated reason: userinfo, query_params, fragment, sensitive_path_segment. |
service_id | string | x402station service identifier from the public catalog surface. |
service_name | string or null | Service display name. |
provider | string or null | Provider/domain grouping used for concentration checks. |
category | string or null | Catalog category when supplied upstream. |
classification | string | Primary class chosen from live, decoy_price_extreme, zombie, dead, dead_7d, mostly_dead, never_paid_zombie, new_provider, slow, price_outlier_high, high_concentration. |
classification_reasons | array or pipe-separated string | All classes that fired for the row. |
first_seen_at | ISO 8601 or null | First time x402station.io saw the endpoint. |
last_probe_at | ISO 8601 or null | Most recent probe included in the aggregate window. |
total_probes_30d | integer | Count of HTTP-naked probes in the trailing 30-day window. |
uptime_30d_pct | number or null | Percent of 30-day probes returning HTTP 200 or 402. |
latency_p50_ms | integer or null | 30-day p50 latency across probes with an HTTP status. |
latency_p95_ms | integer or null | 30-day p95 latency across probes with an HTTP status. |
latency_p99_ms | integer or null | 30-day p99 latency across probes with an HTTP status. |
price_amount_usdc | number or null | Current public catalog price amount, normalized as USDC where supplied. |
price_currency | string or null | Current public catalog currency. |
cdp_l30d_calls | integer or null | Coinbase CDP Bazaar 30-day settlement count when synced. |
cdp_l30d_unique_payers | integer or null | Coinbase CDP Bazaar 30-day unique payer count when synced. |
cdp_last_called_at | ISO 8601 or null | Last CDP-observed settled call timestamp when synced. |
cdp_last_synced_at | ISO 8601 or null | Last CDP quality sync timestamp. |
has_free_trial | boolean or null | Whether CDP enrichment reports a free trial path. |
generated_at | ISO 8601 | Snapshot generation timestamp. |
Probe Methodology
x402station.io probes every indexed x402 endpoint on an approximately 10-minute cadence. A probe is an HTTP naked call: no PAYMENT-SIGNATURE, no paid settlement, and no attempt to consume protected paid content. The probe records aggregate availability evidence such as status class, network failure, and latency.
This public dataset contains endpoint-level aggregates only. It intentionally does not include raw probe-by-probe rows, raw response bodies, payment signatures, wallet identifiers, webhook payloads, or per-request time-series.
Classification Heuristics
The public classes mirror the x402station.io signal vocabulary documented at https://x402station.io/spec#signals.
| Class | Heuristic |
|---|---|
live | No v0.1 dataset class fired. Check total_probes_30d and last_probe_at before treating this as a guarantee. |
decoy_price_extreme | Current price is at least $1,000 USDC. |
zombie | At least 3 probes in the last hour and zero healthy responses. |
dead | At least 3 unhealthy probes in the last 30 minutes. |
dead_7d | At least 20 probes in 7 days and zero healthy responses. |
mostly_dead | At least 20 probes in 7 days and uptime below 50%. |
never_paid_zombie | CDP synced, zero 30-day settled calls, no historical last-called timestamp, endpoint older than 30 days. |
new_provider | Service first seen less than 24 hours before export. |
slow | Last-hour average latency at least 2,000 ms, or 30-day p99 latency at least 5,000 ms. |
price_outlier_high | Current price is at least 10x the provider/domain group median. |
high_concentration | Provider/domain group accounts for at least 5% of active catalog endpoints. |
When multiple classes fire, classification is the highest-priority class and classification_reasons lists the complete set.
Refresh Cadence
The intended public cadence is weekly snapshot refresh. The paid Preflight API remains the fresh path for production agents that need current verdicts, batch checks, and SLA behavior.
License
CC-BY-4.0. Downstream usage should cite x402station.io and this dataset URL.
Citation
Plain text:
x402station.io. x402station.io Preflight Dataset v0.1: aggregated probe data from the x402 agentic-commerce network. Hugging Face Datasets, 2026. https://huggingface.co/datasets/x402station/preflight-dataset-v0_1
BibTeX:
@dataset{x402station_preflight_dataset_v0_1,
title = {x402station.io Preflight Dataset v0.1},
author = {{x402station.io}},
year = {2026},
publisher = {Hugging Face},
url = {https://huggingface.co/datasets/x402station/preflight-dataset-v0_1},
license = {CC-BY-4.0},
note = {Aggregated probe data from the x402 agentic-commerce network}
}
Boundary Policy
Open: endpoint-level aggregate stats, public catalog metadata, classification labels, settlement count/recency aggregates when synced from CDP Bazaar.
Gated: raw probe-by-probe time-series, per-probe response detail, paid API freshness, bulk routing checks, webhook monitoring, and SLA surfaces.
Similar repos
Contributors
3 commits