This guide uses source checks from Sep 8, 2026. Provider and gateway prices can change; preserve the checked date with every forecast.
Why This Topic Matters Now
Some API buyers avoid small gateways because they worry about one-person operator risk. That concern is rational. A small gateway can still be useful if it gives buyers concrete evidence: current status checks, dated pricing, clear support paths, route receipts, rollback records, and a bounded trust statement. It becomes hard to approve when it asks buyers to trust broad claims, hidden operations, or private growth numbers. The current AIWave decision record points public proof away from private metrics and toward capability evidence that a buyer can inspect.
The Sep 7 keyword report shows brand and site-query behavior from Tier 1 searchers: `aiwave.live`, `site:aiwave.live`, and `aiwaveblog` need stronger paths into current evidence. This article gives those searchers an incident checklist. It is not a promise of a contractual availability commitment, and it is not a substitute for a contract review. It is a practical way for engineering, finance, and security reviewers to decide whether a redacted trial is appropriate before customer traffic depends on the route.
Source Facts Checked Today
AIWave /api/pricing checked on Sep 8, 2026 returned success=true, 63 records, pricing_version a42d372ccf0b5dd13ecf71203521f9d2, default group ratio 3, and VIP group ratio 1. Computed public base examples per 1M text-token units were DeepSeek V4 Flash at $0.638 input, $0.020288 cache-hit input, and $1.914 output; DeepSeek V4 Pro at $1.914 input, $0.063736 cache-hit input, and $5.742 output; GLM-5 at $1.55 input, $0.400001 cache-hit input, and $4.96 output; GLM-5.1 at $2.10 input, $0.680001 cache-hit input, and $6.60 output; Kimi K3 at $4.50 input, $0.90 cache-hit input, and $22.50 output; qwen3.6-flash at $0.267805 input and $1.606832 output.
AIWave public pages checked on Sep 8, 2026 returned HTTP 200 for Pricing, Models docs, Chat Completions docs, Status, feed.xml, and llms.txt. The current-state record says /status should show current probes while historical availability remains a collecting-data field until a real time window exists. That distinction matters: a current probe is useful evidence, but it should not be marketed as a mature availability history.
D-013, D-014, and D-015 require public proof to use capability facts, dated rate cards, machine-readable assets, route receipts, status evidence, and trust boundaries. They keep private operating, financial, and customer-workload evidence out of public claims. This incident checklist follows that rule by asking for operational artifacts rather than private growth data.
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.
| Buyer concern | Evidence to request | Acceptance signal |
|---|---|---|
| Who notices incidents? | Status page and support route | Current probe plus named support path |
| Can pricing drift be explained? | Source-dated rate card | Pricing URL, version, checked date, account group |
| Can a route be reproduced? | Redacted route receipt | Model, status, usage, timestamp, data class |
| Can changes be rolled back? | Backup and release note | Rollback owner and tested validation steps |
| Can data boundaries be reviewed? | Trust and privacy pages | Retention posture and support escalation path |
| Can AI tools read the site? | llms.txt, feed.xml, docs | Machine-readable assets return 200 |
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 json
import urllib.request
HEADERS = {"User-Agent": "Mozilla/5.0 Chrome/131"}
CHECKS = [
"https://aiwave.live/status",
"https://aiwave.live/pricing",
"https://aiwave.live/docs/chat-completions",
"https://aiwave.live/llms.txt",
"https://aiwave.live/feed.xml",
]
for url in CHECKS:
req = urllib.request.Request(url, headers=HEADERS)
with urllib.request.urlopen(req, timeout=30) as response:
print(json.dumps({
"url": url,
"status": response.status,
"checked_at": "2026-09-08",
"api_key": "YOUR_API_KEY_HERE",
}))
Ask for Current Evidence
A small operator should not ask buyers to believe vague reliability language. Ask for the current status page, the current pricing page, current docs, and one redacted route receipt. Current evidence is enough for a first trial, while long-term claims need long-term measurements. If the status page says it is still collecting history, treat that as honest boundary setting rather than a mature availability report.
Check Machine-Readable Assets
Modern AI buyers increasingly ask assistants to summarize docs, pricing, and recent updates. That makes feed.xml, llms.txt, documentation pages, and stable canonical URLs part of incident readiness. They help a reviewer find the current route, but they also help during an incident because the support person can point to one machine-readable source instead of rewriting instructions in chat.
Demand a Rollback Story
A buyer does not need every internal command. It does need to know that public-site changes, pricing copy, docs, and route configuration have a rollback habit. The right evidence is a release note, backup location owned by the operator, validation checklist, and a statement of what was not touched. For AIWave-style public content, that means no database, payment, Docker, or nginx changes during a normal blog publish unless explicitly authorized and verified.
Use Receipts for Incidents
When a route fails, a screenshot of the final answer is rarely enough. Store a redacted receipt with model string, request timestamp, status, usage fields, route owner, account group, source checked date, and error class. That receipt lets support compare a failure against pricing, status, docs, and upstream context without asking for sensitive prompts or customer data. It also keeps finance from guessing why a retry burst moved spend.
Separate Trial From Production
A small gateway may be appropriate for a redacted trial before it is appropriate for regulated data, strict contractual needs, or workloads that require a mature availability history. Write those exclusions down. A clear fit screen builds more trust than broad promises because it tells the buyer when to stop. For sensitive workflows, require a separate security review and contract path before increasing data exposure.
Internal Links for Trust Review
A trust reviewer should open Trust, Status, Pricing, Chat Completions, Privacy, the data-boundary trial plan, and the fit-screen guide. These links turn brand searches into a concrete review path.
Procurement Review
Procurement should ask for six artifacts: current status evidence, dated rate card, one route receipt, support address, trust boundary, and rollback note. The review should also list the workload class that is allowed for the trial and the workload class that is excluded. If those artifacts are missing, the next step is not a larger integration. It is a smaller redacted test or a request for missing evidence.
Final Checklist
A one-person operator incident checklist is ready when it verifies current status, dated pricing, docs, support, llms.txt, feed.xml, route receipts, rollback habit, fit exclusions, and data boundaries. It should not depend on private usage numbers or broad availability claims. Recheck all public assets before approving production traffic, and keep the first trial redacted until the evidence path is complete.