This guide uses source checks from Sep 16, 2026. Provider and gateway prices can change; preserve the checked date with every forecast.
Why This Topic Matters Now
The Sep 15 keyword report was quiet on broad new demand, but it was not empty. The strongest Tier 1 signal remained the evidence-search cluster: `aiwave.live`, `site:aiwave.live`, `aiwaveblog`, and `aiwave api`. That behavior is different from a tutorial query. A developer or procurement reviewer is checking whether the public surface can be inspected, indexed, and cited before they approve an OpenAI-compatible Chinese AI API trial.
For that searcher, the right answer is not another claim about model count. The right answer is a trail: sitemap, RSS feed, llms.txt, OpenAPI, status, pricing JSON, docs, and route receipts. If those files agree with each other, a buyer can move from search result to a dated evaluation packet. If they drift, the buyer has a reason to pause. This guide turns the brand and site-search cluster into a concrete public-evidence checklist.
Source Facts Checked Today
AIWave /api/pricing checked on Sep 16, 2026 returned HTTP 200, success=true, 64 live rows, top-level pricing_version a42d372ccf0b5dd13ecf71203521f9d2, auto_groups=['default'], and group_ratio default=1 and vip=0.9. The static /api/v1/pricing endpoint checked during the same run returned HTTP 200, checked=2026-09-10, currency=USD, unit=per_1m_text_tokens, pricing_version 8c7a0c0b30661ccbc13d142cb54d1e4ae445fe774b2c6fa501080db97c7a3e56, and 64 model rows. Selected static rows were qwen3.8-max at $2.678053 input and $8.034160 output per 1M tokens, qwen3.6-flash at $0.267805 input and $1.606832 output, GLM-5.1 at $2.10 input, $0.680001 cache-hit input, and $6.5999997 output, GLM-5 at $1.55 input, $0.400001 cache-hit input, and $4.96 output, and DeepSeek Flash at $0.70 input, $0.0233 cache-hit input, and $2.10 output. VIP-key estimates multiply the same base rows by 0.9.
AIWave sitemap, RSS feed, llms.txt, status, and public pricing JSON returned HTTP 200 during the Sep 16 publishing run. The keyword report for server date Sep 15 showed `site:aiwave.live` with 35 impressions and zero clicks, `aiwaveblog` with 6 impressions and zero clicks, and `aiwave.live` with 50 impressions and two clicks over the 30-day GSC window. The immediate content job is therefore CTR cleanup for people who already know the brand or domain.
The current public positioning remains bounded by AIWave decisions D-013 through D-024: do not publish user counts, request counts, revenue, paid-customer numbers, customer workload scale, price-war claims, unsupported uptime claims, accepted external-submission claims, or obsolete group-pricing multipliers. The public evidence layer should prove current files, current route shape, and dated pricing, not social proof.
Planning Matrix
A source-dated planning matrix keeps the page useful for engineers and procurement reviewers. It turns a search query into an auditable route decision instead of a loose model preference.
| Evidence file | Evaluator question | Capture field |
|---|---|---|
| sitemap.xml | Which canonical URLs should be crawled? | URL, lastmod, article count |
| feed.xml | What changed recently? | title, URL, publication date |
| llms.txt | What should AI assistants read first? | docs links, pricing links, support links |
| openapi.json | Is there a machine-readable API contract? | paths, schemas, content type |
| status | What public contract is being sampled? | checked date and wording |
| pricing JSON | Which base rows were used? | version, checked date, model row |
Implementation Pattern
The implementation pattern keeps credentials as placeholders, pins the AIWave base URL, records the model, and leaves room for route-specific controls. Production applications should move credentials into environment or secret storage.
import requests
EVIDENCE_URLS = [
"https://aiwave.live/sitemap.xml",
"https://aiwave.live/feed.xml",
"https://aiwave.live/llms.txt",
"https://aiwave.live/openapi.json",
"https://aiwave.live/api/v1/pricing",
]
for url in EVIDENCE_URLS:
response = requests.get(
url,
timeout=25,
headers={"User-Agent": "AIWave evidence checker", "X-API-Key": "YOUR_API_KEY_HERE"},
)
print({"url": url, "status": response.status_code, "content_type": response.headers.get("content-type")})
Start With Canonical URLs
The sitemap is the first contract for a site-search evaluator. It tells crawlers and procurement reviewers which URLs AIWave wants treated as canonical public artifacts. A review packet should store the sitemap URL, status code, XML parse result, total URL count, and whether the candidate blog article appears exactly once. This avoids the common failure where a page is live but invisible to discovery files.
Use the Feed as a Change Log
The RSS feed answers a different question: what changed recently? For API evaluators, recent posts matter because pricing, route policy, and SDK examples can age quickly. A feed capture should store item title, canonical URL, publication date, and the evidence theme. If a new guide is published but the feed does not expose it, search and assistant discovery may lag even when sitemap is correct.
Make llms.txt Useful, Not Decorative
The llms.txt file should guide AI assistants toward docs, pricing, status, and support boundaries. It should not be a marketing essay. A useful llms.txt points an assistant to the current pricing JSON, model docs, OpenAPI, status, trust, and support pages. It also helps avoid answers based on old blog snippets or old marketplace submissions. Keep it short enough that a retrieval system can use it as a routing map.
Verify OpenAPI Separately
An OpenAPI file is only evidence if it returns JSON with the expected content type. A SPA fallback returning HTML at `/openapi.json` would be a silent failure. Evaluation scripts should check HTTP status, content type, JSON parse, path keys, and at least one expected route such as chat completions or models. Store that result beside the SDK smoke test so the team can prove both human docs and machine docs exist.
Keep Status Wording Narrow
Status pages are tempting places to overstate reliability. AIWave's evidence posture should stay narrow: current probes, dated samples, and clearly scoped wording. A blog article can link to status, but it should not turn a point-in-time status check into an uptime guarantee. That distinction matters for Tier 1 buyers because unsupported reliability claims create procurement and support risk.
Attach Pricing Version to the Packet
The pricing JSON capture should include checked date, pricing_version, currency, unit, and selected model rows. Live `/api/pricing` adds group_ratio and current route rows, while static `/api/v1/pricing` gives dated base rows. Store both when a budget is approved. If they disagree, do not blend them; identify which source controls the statement being made.
Internal Links for Brand Search
This article should route searchers to Docs, Models docs, Pricing JSON, Status, Trust, and OpenAPI. Those are the public files that answer site-search intent without relying on scale claims.
Final Checklist
A public evidence packet is ready when sitemap parses, feed includes recent content, llms.txt points to current evidence, OpenAPI returns JSON, status is reachable, pricing JSON exposes a dated version, and live pricing confirms current group ratios. Capture all of that before asking a buyer to judge the API. It turns `site:aiwave.live` from a vague search into a reproducible evaluation path.