01
Prefer the source owner
Provider documentation and announcements are the default authority for price, limits, access and lifecycle claims.
Home / About & methodology
About VerticalAPI
Decision intelligence for a market that changes every week.
VerticalAPI helps developers and technical buyers choose an LLM API using current, source-checked evidence—not an affiliate list, an opaque score or a signup funnel.
50model pages
61comparisons
25provider profiles
173sitemap URLs
Repository inventory snapshot · verified 2026-08-23 · these are coverage counts, not customer metrics
Why this exists
Choosing an API should not require reconciling ten tabs.
Model names, list prices, context limits, access routes and lifecycle states move independently. Comparison pages often copy a number once, remove its conditions and leave it online after it changes.
VerticalAPI organizes those facts around the buyer’s decision: which model or provider fits this workload, what does it cost now, and what evidence supports the answer?
The product is the connected corpus—model pages, provider profiles, comparisons and benchmarks—plus the discipline to keep uncertainty visible.
Methodology
Four rules for every decision page.
Different evidence types answer different questions. We keep them separate rather than compressing them into a single unexplained score.
02
Date facts that drift
Pricing, model availability and rate limits carry a checked date and relevant conditions.
03
Separate measured from claimed
VerticalAPI measurements include method and limitations. Provider benchmark claims remain attributed to the provider.
04
Keep gaps visible
If a field is unavailable or not published, it stays incomplete. Missing evidence is not zero performance.
How recommendations are formed
A recommendation begins with explicit workload constraints—price, latency, context, modality, geography or operational control. A model is not declared “best” in the abstract. Where the current evidence cannot decide, the page says so.
Change cadence
Factual errors are corrected as soon as a reliable source is available. Structural changes to established comparison pages are paced and reviewed so freshness work does not destroy useful search coverage.
Corrections
A stronger source wins.
Submit evidence
Found a price, model status or provider fact that changed?
Send the page URL, the exact field and a first-party source. We review factual corrections without charge. Sponsorship, partnership or customer status does not affect the outcome.
Request a correctionCommercial disclosure
Sponsored placements, if offered, are labelled in place. They cannot change the factual record, methodology, editorial ranking or conclusion. Custom research deliverables distinguish verified facts, analysis and unknowns.
Contact
Use the structured request route for data corrections, provider listings and research, or email [email protected] for security disclosures. The machine-readable site guide is available at /llms.txt.
FAQ
Boundaries that matter.
Is VerticalAPI a model provider?
No. VerticalAPI does not train or host foundation models. It publishes decision data about models and API providers; a separate beta gateway can route requests to upstream providers.
Is every field independently measured?
No. Provider facts such as list price and documented limits are sourced from the provider. VerticalAPI measurements are explicitly labelled and include methodology. The two are not presented as the same evidence.
Does a listing imply endorsement?
No. A profile records available evidence. Inclusion does not mean VerticalAPI recommends the provider for every workload.
Who builds VerticalAPI?
VerticalAPI is built by Brandyze in Aix-en-Provence, France.