// knotbind gateway

One endpointfor every AI model.

Your apps call one OpenAI-compatible URL with a Knotbind access key. Each key has a monthly budget and a request rate. If a provider is down, a fallback list is a separate Model Failover module, not part of this plan.

Offered to pilot customers. Plans and prices may change after the pilot.

// what you get

A budget on every key. One URL to change.

Budgets and keys

Give each app its own access key with a monthly budget and a request rate. Your provider keys stay sealed in the workspace, never in the app.

Usage on the key

Each key reports its request count, success rate, latency, tokens and cost. A workspace-wide split by provider and model is the Request Analytics module.

Errors an agent can use

Opt in per key. Hints use the HTTP status and validation fields the provider already named. They do not invent a renamed field or retry a write.

Modules, when you need them

Five optional add-ons, off until you enable them. See the Gateway modules.

// gateway modules

Optional controls on the Gateway.

A Pro or Business plan already includes the endpoint and per-key budgets. These add-ons are off until you enable them. Prices are on the pricing page; the problem each one solves is on the services page.

Semantic Cache

Cache LLM responses by exact match or semantic similarity. Cut costs and latency for repeated or similar prompts.

Content Guardrails

Redacts email addresses, US phone numbers, US Social Security numbers and payment card numbers in JSON chat message and reasoning text.

Request Analytics

Groups recent Gateway requests by model and provider on the Gateway page.

Model Failover

A fallback model chain per access key. Enabling it creates no rules until you add one.

Advanced Rate Limits

An extra per-key request quota. Enabling it creates no quota until you add one.

// switching over

Three steps, usually one afternoon.

Add your providers

Tell us which model providers you use. Keys are stored encrypted.

Set a budget

Create an access key with a monthly budget and a request rate. Add a failover rule only if you buy that module.

Change one URL

Point your app's base URL at your Knotbind Gateway endpoint.

// pricing

Included in Pro and Business.

Model usage is billed by your providers, on your provider key. Budgets are included with the Gateway. Cache, failover, redaction and workspace analytics are priced separately.

Knotbind Gateway is included in Pro ($199 per month) and Business ($399 per month).

See the plans

// questions

Gateway questions.

What is the Knotbind Gateway?

One OpenAI-compatible endpoint between your apps and the models they call. Most apps switch by changing a base URL. Pro and Business include an access key per app, a monthly budget, a request rate and per-key usage. Caching, model failover, response redaction and workspace-wide analytics are optional modules.

Which models can it route to?

OpenAI, Anthropic, Google, Mistral, AWS Bedrock, Azure and other major providers, plus models you host yourself. You bring your own provider keys.

Does the Gateway include fallback and caching?

No. Model Failover ($29 per month) adds a fallback list for one access key after you create the rule. Semantic Cache ($19 per month) answers repeated or similar prompts from cache. Neither runs just because the Gateway is on.

// contact

Tell us what your customers should be able to ask.

Send a message. We reply with what to connect, which assistants to target, and a fixed price. No call.