
SproutAI Gateway deploys inside your private network and routes requests across AWS Bedrock, Azure AI Foundry, OpenAI, and Anthropic — with real-time app-level budgeting and a path to enterprise Azure Purview governance.
Connect Bedrock, Azure AI Foundry, OpenAI, and Anthropic through one unified spec. Switch providers without touching application code.
Set monthly dollar limits and token quotas per app or team. Spend updates live, with guardrails before bills surprise you.
Partner with Sproutlogix on tailored Azure Purview integrations, PII masking, and automated audit catalog syncs.
Connect AWS Bedrock, Azure AI Foundry, OpenAI APIs, and Anthropic APIs through a unified API spec. Switch models dynamically without changing app code.

Eliminate vendor lock-in, monitor multi-model usage, and enforce strict spending guardrails across all AI endpoints.
Set dynamic monthly budget limits per application, business unit, or developer team.
Cap token consumption to prevent cloud billing surprises.
Automated fallback routing keeps requests flowing when a limit is hit.
Tailored Azure Purview integrations built by Sproutlogix consultants.
Custom PII classification and masking rules for sensitive data.
Automated audit pipelines synced into your corporate Purview catalog.
Email key owners at 80%, 90% and 100% of budget or token limit, with per-key thresholds and daily usage digests.
Filter and group cost and tokens by team, owner, key, model or date, then export to Excel or CSV.
Admin and user roles with no external identity provider. Users only see usage for the keys they own.
Point Claude Code, aider, llm or LangChain at the gateway. Claude models pass through natively, with streaming, tool use, thinking and prompt caching. The console generates a ready-to-paste setup from your model routes.
export ANTHROPIC_BASE_URL="https://your-gateway/anthropic"
export ANTHROPIC_AUTH_TOKEN="<your-gateway-api-key>"
claudeKeep using the official OpenAI and Anthropic SDKs and change only the base URL and key. The console includes copy-ready Python, Node.js and cURL samples with your model names filled in.
client = OpenAI(
base_url="https://your-gateway/openai/v1",
api_key=os.environ["GATEWAY_API_KEY"],
)Core gateway capabilities are 100% free. Enterprise compliance setup is tailored by Sproutlogix experts.
| Feature / Capability | Free Community Edition | Sproutlogix Consultancy |
|---|---|---|
| Self-Hosted Private Deployment (Docker/Helm) | ✓ Included | — |
| AWS Bedrock & Azure AI Foundry Connectors | ✓ Included | — |
| Direct OpenAI & Anthropic API Routing | ✓ Included | — |
| App-Level Budget Hard-Caps & Token Tracking | ✓ Included | — |
| Claude Code, CLI & OpenAI/Anthropic SDK Support | ✓ Included | — |
| Usage Alerts, Daily Summaries & Excel/CSV Reports | ✓ Included | — |
| Built-in Login with Admin & User Roles | ✓ Included | — |
| Azure Purview Lineage & Audit Integration | — | ★ Specialized Service |
| Custom Gateway Plugin & Middleware Engineering | — | ★ Specialized Service |
| Internal SSO / OAuth2 Integration | — | ★ Specialized Service |
Deploy unlimited gateway nodes inside your private environment. Apache 2.0 licensed and free for commercial use.
Expert engineering to connect your gateway with enterprise compliance and cataloging tools.
All four primary ecosystem integrations — AWS Bedrock, Azure AI Foundry, OpenAI APIs, and Anthropic APIs — are fully enabled in the Community Edition, with no artificial throughput caps or usage paywalls.
Enterprise Purview deployments require custom data classification models, PII taxonomy mapping, and tenant-specific permissions. Sproutlogix solution architects deliver an integration tailored to your organization's compliance standards.
No. SproutAI Gateway is deployed entirely within your infrastructure — AWS VPC, Azure VNet, or on-premises Kubernetes. Zero telemetry or prompt data ever leaves your perimeter.
Yes. Claude Code, aider, llm, LangChain and the official OpenAI and Anthropic SDKs all work by pointing them at the gateway's URL with a gateway API key. Claude models are passed through natively, and every request is counted against that key's budget and shows up in usage reports.
Docker and a MySQL database. The API and management console ship as a single container, and the included docker-compose file runs MySQL for you. For production you can use your own MySQL (such as Amazon RDS or Azure Database for MySQL) and run several gateway replicas behind a load balancer.
Ready to unify and govern your enterprise AI infrastructure? Contact us for a demo or a consultation.
Six questions, two minutes, no signup. See how your data readiness compares to published benchmark data.