What vechgate actually is?
You will never got worry about token limit anymore
Vechgate is an AI gateway that abstracts provider and model complexity while keeping routing and billing behavior visible. Ordinary requests route automatically. Specialized requests pay for what they ask for.
01 / how it works
Every request follows one inspectable pipeline
The gateway validates, measures, routes, executes, and records. No step is hidden from the usage log.
- 01 authenticateThe API key is validated before any provider work starts.
- 02 limitsRPM and concurrency are checked. Over the limit is HTTP 429 with retry metadata.
- 03 inspectThe request is inspected for complexity, capabilities, and context size.
- 04 overrideIf a model override is present, the premium entitlement is checked, the surcharge estimated, and spending limits verified.
- 05 selectOtherwise the router selects the lowest-cost eligible model that meets the request requirements.
- 06 executeInference runs against the chosen provider.
- 07 recoverA failure triggers retry or fallback within the request policy, never a silent model substitution.
- 08 recordAn immutable usage record is written with the actual model, latency, and price applied.
02 / principles
Rules the product is built around
These are operating rules, not aspirations. They hold on every request.
automatic_by_defaultAutomatic routing is the default mode for standard requests.
no_silent_swapA premium override is never quietly downgraded to automatic routing; an unmet override fails clearly.
rpm_is_the_capRPM caps the account. There is no per-token customer price.
auditable_billingEvery charge traces to an immutable usage record, and corrections use compensating entries.
operate_liveAdministrators can disable a provider or model without redeploying the gateway.
03 / roles
Who uses Vechgate, and what they can do
Access is scoped to an organization from the first login.
guestReads the product and pricing.
developerCreates API keys, sends requests, and views usage.
account_ownerManages billing, subscription, and team access on a premium plan.
team_memberUses assigned keys and views permitted usage inside one organization.
adminManages providers, models, pricing, routing, and accounts.
systemPerforms routing, billing, monitoring, and automated failover.
Read the pricing, then try the endpoint
RPM tiers start free. Premium model overrides are a fixed surcharge per request on paid plans.