Integrations · developer clients

Governed access, in the tools they already use

Developers keep their workflow. AIMS meets them inside the editor, the CLI, and Copilot, with entitled models, no key management, and attribution stamped on every request.

Short-lived keysClient-owned providersFeature-level attribution
01 · Developer clients

Where your developers already work.

One governed gateway. Every familiar workflow.

VS Code extension

Connected

Use AIMS directly from the native chat workflow, with project and feature attribution built in.

Synapx AIMSNew chat
YouRefactor the checkout retry logic and add tests.
Synapx AIMS · GPT-4.1
GPT-4.1 · entitled1.2k / 5k tokens
  1. VS Code Chat
  2. AIMS Gateway
  3. Policy Check
  4. Entitled Model

GitHub Copilot BYOK

Connected

Govern Copilot through short-lived credentials and entitled models.

Copilot
1.2ktokens
Copilot(BYOK client)AIMS GatewaySecure trust boundary

Press send to route a request through AIMS.

Key expires in 3:58:21

CLI

Connected

Run coding agents through the governed gateway without storing provider keys.

checkout-redesignGPT-4.1714 ms
  1. CLI
  2. AIMS Gateway
  3. Policy Check
  4. Secure Execution
  • Familiar workflowsUse the tools your developers already know.
  • Built-in governanceIdentity, policy, and entitlements.
  • Consistent attributionEvery request carries project and feature context.
02 · Live infrastructure

AI traffic control centre

Every request passes through one control plane before reaching a model.

Incoming requests
  • VS CodeNative extension1,370 / min
  • GitHub CopilotBYOK endpoint1,240 / min
  • CLISecure process685 / min
  • JetBrainsComing soonPreview
  • API SDKComing soonPreview
Illustrative traffic · per minute
AIMS GatewayLive
OperationalGoverned routing
  • Entitlement EngineActive
  • Policy EnforcementActive
  • Budget ControlsActive
  • AttributionActive
Short-lived credentialsIssued just in timeTelemetryAttributed per request
Routed to models
  • GPT-5EntitledEntitled · via OpenAI1,420 / min
  • ClaudeEntitledEntitled · via Anthropic980 / min
  • GeminiEntitledEntitled · via Google540 / min
  • Other OpenAIProvider routeYour OpenAI org key240 / min
  • Azure OpenAITenant routeYour Azure tenant115 / min
3,295 / minRequests routed
4 of 4Checks on every request
12Requests blocked by policy

03 · Security by design

Central control. No long-lived keys

AIMS keeps provider credentials behind the gateway while developers work where they are fastest.

A developer's request passes through the AIMS Gateway, which authenticates it, scopes it, and issues a short-lived credential for project-x, feature-x, and claude-sonnet. AIMS uses it to reach the AI provider, and the credential is revoked automatically when its time to live ends.

Credential control plane

Live
Developer Uses familiar tools (IDE, CLI, Copilot)
AIMS Gateway Central control plane
  1. AIMS confirms who is asking and which project the request belongs to.
  2. Access is limited to the project, feature, and models the developer is entitled to.
  3. A short-lived credential is created at request time, tied to this request's context.
  4. AIMS calls the provider within that scope. Provider keys stay behind the gateway.
  5. Access ends when its window closes, or at once when it is revoked centrally.
AI Provider Access via short-lived credentials

See the full security model

04 · How it works

From provider to productive session

Connect once, then let AIMS keep the governance record current while developers get on with the work.

1

Connect your provider

Bring your Azure AI Foundry, Azure OpenAI, or other provider account to the encrypted AIMS Gateway.

2

Import your board

Turn GitHub Projects items into AIMS features and cost centres for delivery.

3

Develop in your tools

Use VS Code, Copilot BYOK, or the CLI with entitled models and no extra workflow.

4

Spend attributed automatically

Every request is tied to its project and story, with budgets and evidence kept current.

05 · Providers & models

Your providers, your models, your board

Choose the capability, control who can use it, and make every model decision a costed decision.

Your providers

Bring your own provider, in your own tenancy.

  • Azure AI FoundryEnterprise ready Connected
  • Azure OpenAIYour tenant Connected
  • AnthropicDirect connection Coming soon
  • OpenAIDirect connection Coming soon

Keys stay behind the gateway

Your models

A catalogue you can compare.

Around 400 models in the full catalogue

Your board

Turn work into cost centres.

GitHub ProjectsImport and sync

Synced from GitHub Projects hourly

Costed decision

Every request is attributed.

Current spend · this sprint
£1,284.20
86% of £1,500 budget Near limit
Project
checkout-redesign
Feature
cart-flow
Model
GPT-4.1
Cost / request
£0.42
Status
Entitled Not entitled

  1. Providers Connected and secure
  2. Models Compare and choose
  3. Board Link your work
  4. Costed Automatically attributed

The AIMS difference

Make every AI request accountable.

Govern access, attribute spend, and keep developers in the tools they already use.

Request early access