AuspexIQ / README.md
pima5's picture
video_context free tier
cd27f43 verified
|
Raw
History Blame Contribute Delete
2.81 kB
---
title: AuspexIQ
emoji: "๐Ÿ”ฎ"
colorFrom: indigo
colorTo: purple
sdk: docker
app_port: 8000
pinned: false
---
# AuspexIQ
Reads the signs before you launch. AuspexIQ is a live analysis service that tells
any calling agent whether a YouTube niche is worth entering โ€” using the
YouTube Data API v3 at request time to find outlier videos, measure
saturation, and return a verdict: **ENTER / CROWDED / AVOID**.
No mock data, no cached opinions, no hallucinated stats. Every number in every
response comes from a live API call (or a short-TTL cache of a previous live
response). Priced per call.
## Endpoints
Per-call pricing on X Layer (OKX Agent Payments Protocol).
Calling without payment returns `HTTP 402` + `PAYMENT-REQUIRED`;
x402-capable agents pay and retry automatically:
- `GET https://pima5-auspexiq.hf.space/paid/scan_niche?query=...` โ€” 0.1 USDT
- `GET https://pima5-auspexiq.hf.space/paid/channel_outliers?channel=...` โ€” 0.1 USDT
- `GET https://pima5-auspexiq.hf.space/paid/video_context?video=...` โ€” FREE
- `GET https://pima5-auspexiq.hf.space/paid/rising_channels?niche=...` โ€” 0.25 USDT
- Health: `https://pima5-auspexiq.hf.space/healthz` (free)
Available on OKX.AI as Agent #9453.
## Tools
### `scan_niche`
Assess a niche keyword: who ranks, how concentrated it is, which videos are
outliers relative to their own channel's baseline, and whether a new channel
should enter.
| Input | Type | Default | Notes |
|---|---|---|---|
| `query` | string | required | niche keyword, 2โ€“80 chars |
| `region_code` | string | `US` | ISO 3166-1 alpha-2 |
| `recency_days` | int | 365 | 30โ€“1825; only videos in this window |
| `max_results` | int | 50 | 10โ€“50 search results analyzed |
Returns `saturation_score` (0โ€“100), `verdict` (`ENTER` / `CROWDED` / `AVOID`)
with number-citing reasons, signals (channel diversity, top-3 view
concentration, 90-day freshness, small-channel breakouts), and up to 15
outlier videos โ€” each with a watchable URL, its channel's baseline, and the
outlier multiple.
### `channel_outliers`
Reveal which of a channel's recent videos overperformed the channel's own
baseline (median views of recent long-form uploads).
| Input | Type | Default | Notes |
|---|---|---|---|
| `channel` | string | required | `UC...` ID, `@handle`, or channel URL |
| `lookback_videos` | int | 30 | 10โ€“100 recent uploads |
| `min_multiple` | float | 2.5 | 1.5โ€“10, views/baseline threshold |
## Error model
Failures return a structured error, never a fabricated result:
```json
{"ok": false, "error": {"code": "MISSING_API_KEY | QUOTA_EXHAUSTED | YT_API_ERROR | CHANNEL_NOT_FOUND | INVALID_INPUT | UPSTREAM_TIMEOUT", "message": "...", "retryable": true}}
```
Every response's `meta` reports cache state, fetch time, and quota units
spent/remaining for the day.