Sign in
Skip to content

Enterprise AI Gateway

AI within reach.Risk in plain sight.

Create your organization now and get up to 50% off.

Reach the major models through one endpoint, authorize access team by team, control quota in real time, and keep a trace of every call.

  • Granular permissions
  • Live quota control
  • One endpoint
  • Full request trail
New API
Every provider reaches your team through one governed endpoint.

Bring your own workflow

Use New API from the coding tools your team already opens.

Point compatible clients at New API once. Your team keeps its preferred interface while requests travel through the same managed access path.

VS CodeEditor-based AI workflows
CodexAgent and terminal workflows
Claude CodeAnthropic-compatible workflows
TraeAI-first editor workflows
ZcodeCoding assistant workflow
Compatible clientsOpenAI-compatible configuration
New APIManaged endpoint
Authorized model responseReturned to the tool your team chose

Built around one managed route

From the model you choose to the work that reaches users.

New API keeps model choice, team workflows, and delivery paths connected through one operating layer.

01

Approved models

Give teams useful choice without multiplying integrations.

OpenAIgpt-4o
AnthropicClaude
GoogleGemini
More approved models
02

Team workflows

Use AI where work already happens, from experiments to review.

Product prototypes
Engineering automation
Operational review
Shared team work
03

Delivery paths

Let every approved user begin from the right interface.

Compatible API
Team workspace
Managed API keys
Request records
New APIOne managed AI gateway

For everyone else

Not everyone on your team writes code.

Employees sign in and start working straight away. The chat workspace is part of the gateway, not a separate product to buy: no keys to copy, no SDK to install.

  • Nothing to installA browser and their own account is all an employee needs.
  • Only the models you authorizedThe same permissions that govern your API traffic apply here.
  • Counted the same wayEvery conversation lands in the same usage, quota and cost reporting.
claude-opus-5

Summarize this quarter’s support tickets by theme.

Ask anything, or paste a document

One AI gateway for the work your teams already do.

New API is an enterprise AI gateway for teams that need one governed way to use multiple AI models. It keeps AI API access useful for builders and visible to the people responsible for access and spend.

A unified AI API for multiple models

Use one familiar API integration to reach the models your organization authorizes. Teams can change the model parameter instead of rebuilding an integration for every provider.

AI API management with practical controls

Set model access by team or member, issue and manage API keys, and apply quotas before unapproved or over-limit requests reach a provider.

AI usage and cost management you can trace

Request logs record the caller, model, status, and token usage. That gives engineering, finance, and security teams a shared record for operations and review.

Simpler to adopt, and fully recorded.

Change the model, not the code.

One request format reaches every major model. Change a single parameter to switch, with no business logic to rewrite.

A chat request where only the model parameter changes between providers.

Every call leaves a record.

Time, caller, model, status, and token spend are all recorded, ready to trace or export at any time.

TimeAPI keyModelStatusTokens

Getting started

Questions before your first AI workflow.

A practical starting point for teams deciding how to introduce New API into real work.

Where should a team begin?

Start with one workflow that already has a clear owner, such as a product prototype, an internal assistant, or a development integration. Establish the access path and review the early usage before expanding.

Do we need to replace existing model integrations first?

No. Teams can introduce New API for a new workflow or move integrations incrementally. A compatible API gives developers a familiar way to route approved requests through the gateway.

Can colleagues who do not write code use it?

Yes. The workspace provides an approved place for non-technical teammates to work with AI without needing to manage an API integration themselves.

How do we choose a model for a workflow?

Begin with the capability, quality, latency, and cost expectations of that workflow. Teams can compare approved choices while keeping the calling pattern consistent.

Can New API work alongside our current tools?

Yes. New API is designed to become the managed AI access layer around the workflows and tools your team already uses, rather than requiring one new way to do every task.

Which team should try it first?

Choose a team with a concrete AI use case and someone responsible for the result. Early feedback from that team makes the next rollout decisions more useful for everyone else.