RESEARCH DESK & INDEXING QUEUE
Submit an API, Model, or Pricing Update
Help our research engineering team maintain zero-error telemetry accuracy. Submit newly launched models, report provider price drops, or request developer dataset partnerships.
Provider Listing & Verification Criteria
To ensure CheckAPICost remains the gold standard for enterprise architectural planning and cloud financial engineering, every service indexed on our platform must satisfy strict verification criteria:
- Transparent, Public Rate Cards: We do not index providers with opaque "Contact Sales" barriers or undisclosed minimum spend commitments without verifiable enterprise rate tiers.
- Standardized Billing Metrics: Pricing must be reducible to normalized unit metrics: per million tokens, per gigabyte of egress, per minute of audio, or per SMS segment.
- Documented SLA & Uptime: Hosted models and APIs must demonstrate sustained 99.9% uptime across production availability zones.
- Reproducible Tokenization: Foundation models must specify exact tokenization vocabularies (e.g. cl100k_base, o200k_base, Llama-3-Tokenizer) to ensure fair price-per-token equivalence.
Enterprise Data Feeds & FinOps Syndication
Cloud procurement organizations, venture capital researchers, and FinOps platforms utilize CheckAPICost telemetry to benchmark internal cloud commitments, audit vendor invoices, and project compound AI workload costs.
If your organization requires programmatic JSON feeds, custom automated webhook sync, or historical tariff change archives, contact [email protected] with details on your technical integration requirements.
Frequently Asked Questions
What is the typical response SLA for rate discrepancy reports?
Rate discrepancy submissions and price reduction reports are reviewed by our research engineers within 24 hours of receiving verifiable documentation (official OpenAPI endpoints or vendor rate announcements).
Can cloud providers or AI labs pay for expedited indexing?
No. We do not accept paid indexing or sponsored placement. All model submissions undergo identical objective mathematical validation and synthetic latency testing.
How can enterprise FinOps teams license programmatic data feeds?
Enterprise cloud procurement and FinOps teams can consume machine-readable rate specs via llms.txt and llms-full.txt or contact [email protected] for custom JSON feeds and automated webhook sync integrations.
What criteria must a new API meet to be indexed on CheckAPICost?
Services must have public, non-gated pricing documentation, a documented production API endpoint (REST/gRPC/WebSocket), sustained 99.9% uptime, and standardized unit billing metrics.
How do researchers submit benchmark latency evaluation data?
AI researchers can submit latency datasets including sample sizes (n>=100 requests), standard deviations, region of origin (e.g. AWS us-east-1), and token count distributions by attaching benchmark JSON logs to our research inbox.
Do you provide custom architectural cost modeling consultations?
Yes, our engineering team regularly assists high-throughput AI startups and enterprise FinOps teams with multi-model compound routing topologies, prompt caching optimization audits, and cloud egress reduction plans.
Research Desk SLA & Support Tier Response Times
CheckAPICost prioritizes data accuracy above all else. Inquiries submitted through our research queue are triaged by systems engineers and evaluated according to strict response SLAs:
| Inquiry Category |
Target SLA |
Verification Action |
| Provider Price Reductions |
< 24 Hours |
Automated scraper validation, index synchronization, changelog deployment |
| New Foundation Model Launches |
< 48 Hours |
Synthetic latency probing, benchmark ingestion, model matrix listing |
| Enterprise FinOps Data Feeds |
< 24 Hours |
Custom JSON schema provisioning, webhook endpoint registration |
| Security Disclosures |
< 4 Hours |
Immediate security engineering review and PGP-encrypted triage |
Benchmarking Submission Protocol for AI Labs
AI laboratories and independent hosting providers seeking to submit benchmark figures for inclusion in our 59-Model Matrix must provide reproducible telemetry artifacts:
- Sample Sizes: Latency benchmarks must consist of at least 100 consecutive requests recorded during normal peak availability windows.
- Standardized Payloads: Probes must specify prompt length (standard 1,000 prompt tokens) and generated completion length (standard 256 tokens).
- Tokenizer Verification: Model submissions must indicate the exact BPE tokenizer (e.g. tiktoken cl100k, o200k, or HuggingFace tokenizer.json) to eliminate token count inflation.
To review our existing calculators and model matrices before submitting, explore our 59-Model Matrix, test the Cursor vs Cline Calculator, or examine our Prompt Caching ROI Tool.
Vulnerability Disclosure & Security Protocol
CheckAPICost takes developer infrastructure security seriously. If you identify an operational anomaly, unintended data exposure, or security vulnerability in our client-side calculators or edge proxies, please report it immediately:
- PGP Encrypted Reporting: Send detailed vulnerability descriptions and reproduction steps directly to [email protected].
- Coordinated Disclosure: We adhere to standard 90-day coordinated vulnerability disclosure windows to allow remediation before public reporting.
- Recognition & Bug Bounty: Valid security reports that safeguard developer privacy or calculator integrity are eligible for our public Hall of Fame acknowledgment and engineering honorariums.
Engineering Inquiries, Benchmark Submissions & Technical Corrections
Our research and editorial team comprises cloud architects, systems engineers, and AI infrastructure specialists. We actively welcome verifiable benchmark datasets, corrections on recent provider rate revisions, and collaborative architectural case studies from engineering teams operating large-scale production fleets.
Communication Channels & Response SLAs
| Inquiry Type |
Recommended Channel |
Required Information |
Typical SLA |
| Rate Card Updates / Pricing Corrections |
GitHub Pull Request / Issue |
Link to official vendor docs / commit |
< 24 business hours |
| Hardware / Benchmark Submissions |
Direct Editorial Email |
Reproducible script + latency trace logs |
2-3 business days |
| Enterprise TCO Consulting |
Advisory Contact Form |
Architecture overview & monthly token scale |
1 business day |
Responsible Security Disclosure
If you identify a security vulnerability or anomalous behavior on CheckApiCost web properties, please report it responsibly to [email protected] with detailed reproduction steps prior to public disclosure.
Enterprise Architecture Audits & Custom TCO Modeling
Beyond our public calculators, CheckApiCost provides custom architectural review engagements for high-volume enterprises spending over $10,000 monthly on cloud and AI APIs. Our advisory team reviews architecture diagrams, inspects VPC egress transit routing, audits database pooling configurations, and evaluates LLM prompt caching hit rates, delivering actionable optimization roadmaps that typically reduce cloud expenditures by 35% to 60% within 30 days of implementation.
What information should be included when submitting benchmark trace logs?
Please include the exact target model ID, provider region, prompt token length, generated token count, streaming TTFT timestamp, and reproducible curl command or Python script. Submissions with full HAR network captures receive prioritized verification.
Direct Technical Briefings & Architecture Reviews: We provide dedicated architectural guidance and comprehensive infrastructure spend breakdowns. For venture capital firms, corporate engineering leaders, and hedge funds conducting cloud spend due diligence, our research team offers direct 30-minute technical briefings on emerging AI hardware unit economics.