Overview
UNIFIED MULTI-PROVIDER GATEWAY Route requests to Anthropic Claude, OpenAI, AWS Bedrock, Google Vertex AI, Azure OpenAI, and Google GenAI through a single API. Full Claude API and OpenAI API compatibility - switch providers without code changes. Built-in health checks, automatic failover, and model discovery.
MCP GATEWAY Connect MCP (Model Context Protocol) servers once and expose their tools to all users and agents. Supports Global and Personal servers with secure ECDSA key exchange, real-time tool discovery via SSE, heartbeat monitoring, and encrypted credential storage. Agents integrate once and access all enabled tools.
INTELLIGENT CACHING Automatic prompt-level caching with adaptive boundary calculation eliminates redundant LLM calls. Cache profitability analysis applies caching only when cost-effective. Supports Bedrock-native cache TTLs. Tracks hit rates, miss costs, and generates per-user Cache Hero Scores to gamify optimization.
12-LAYER ANALYTICS ENGINE Basic: requests, tokens, spend, latency, errors, success rates. Personal: token ratios, latency experience, cache efficiency. Team: member performance, budget burn, model mix. Organization: cross-team ROI, vendor risk. Security: API key hygiene, jailbreak detection, anomalies. FinOps: context saturation, cache miss costs, spend forecasting. Agentic: conversation depth, tool chaining, agent failures. Prompt: reusability, compression, complexity. Efficiency: session productivity, hours saved. Session Quality: context-aware scoring. RAG/DX: error impact, rate limit forecasting. Real-Time: WebSocket live dashboards.
COST OPTIMIZER WITH AI RECOMMENDATIONS Simulate rate limits and budget caps against historical usage before applying them. AI-powered optimization suggestions via Claude. Preset configurations for Anthropic and OpenAI tier limits. Detailed breakdowns by model, user, and time period.
PROMPT EVALUATOR AND PROMPT MENTOR 6-dimension scoring: Clarity, Context, Structure, Output Specification, Reasoning, Safety - with letter grades A+ through F. AI-powered rewrite suggestions and best practice detection. Interactive Prompt Mentor for real-time guidance. Team and organization leaderboards.
SKILLS HUB - PLUGIN MARKETPLACE Create, version, publish, and distribute reusable AI skill plugins. Git-based distribution with YAML metadata. Public and private visibility, per-API-key activation, team sharing. AI-generated metadata. Direct Claude Code plugin integration. Build once, deploy across the organization.
6-LEVEL ROLE-BASED ACCESS CONTROL Permission hierarchy: Personal (L1), Team (L2-L3), Department Lead (L4), Org Admin (L5), Super Admin (L6+). Controls analytics access, model tiers, admin functions, spending visibility. Supervisor impersonation in read-only and write mode with full audit logging of every session and action.
ENTERPRISE SECURITY AES-256-GCM encryption for all credentials and API tokens at rest. Zero-knowledge admin access - admins see only masked tokens. Secure token separation: real tokens for auth, analytics tokens for logging. Per-user and per-key rate limiting with auto-unblock. CROWD SSO and Google OAuth.
TEAMS AND ORGANIZATIONS Role-based teams with owner, admin, member, and viewer roles. Share skills, budgets, and resources across teams. Department leads and org admins get aggregated views. Global View for super admins shows metrics across all users.
API KEY MANAGEMENT Named keys with per-key rate limits and budget caps. Key rotation preserves full analytics history. Admin key management via impersonation mode. Deactivation, reactivation, and lifecycle tracking.
FILE STORAGE AND SHARING Project-based file storage with sharing and access control. File overwrite preserves sharing permissions. Auto-registration of users from corporate directory. Direct browser access via web URLs.
SESSION AND CODE METRICS Automatic session management with duration tracking and quality scoring. Git diff integration tracks AI vs human lines per session. Jira task association for development productivity reporting.
REMOTE DESKTOP INTEGRATION WebSocket relay with Noise Protocol encryption (Noise_IK_25519_ChaChaPoly_BLAKE2s). Session sharing, collaboration, and remote UI through secure HTTP tunnels.
FEEDBACK AND RELEASE NOTES Kanban-based feedback workflow for suggestions, bugs, and improvements. AI-generated release notes from completed items using Claude Haiku. Soft-delete archival for audit trail.
AWS MARKETPLACE NATIVE Built-in subscription management, customer lifecycle tracking, metering, and billing event integration.
Highlights
- Single gateway for 6+ LLM providers and unlimited MCP servers. Route requests to Anthropic Claude, OpenAI, AWS Bedrock, Google Vertex AI, Azure OpenAI, and Google GenAI through one API. Full Claude API and OpenAI API compatibility - switch providers without code changes. MCP Gateway connects tool servers once and exposes them to all users with ECDSA key exchange, real-time discovery, and encrypted credentials. Integrate once, access everything.
- 12-layer analytics engine covering every operational dimension - from basic usage, spend, and latency to security anomaly detection, agentic workflow analysis, FinOps cost attribution, prompt quality scoring, and real-time streaming via WebSocket. Team and org-level views with budget burn tracking, model mix analysis, vendor risk assessment, and cross-team ROI. Cache efficiency metrics with per-user Cache Hero Scores gamify optimization across the organization.
- AI-powered cost optimization and prompt quality as measurable metrics. Cost Optimizer simulates rate limits and budget caps against historical data before applying. Claude-generated recommendations with presets for Anthropic and OpenAI tiers. Prompt Evaluator scores prompts across 6 dimensions (Clarity, Context, Structure, Output Spec, Reasoning, Safety) with A+ to F grades, rewrite suggestions, and team leaderboards. Enterprise security: AES-256-GCM encryption, 6-level RBAC, and audit logs.
Details
Features and programs
Financing for AWS Marketplace purchases
Pricing
Dimension | Description | Cost/month |
|---|---|---|
LUMI Enterprise Subscription | Enterprise subscription for the LUMI platform with full platform functionality. | $8,000.00 |
Dimensions summary
Top-of-mind questions for buyers
Vendor refund policy
Buyers may request a refund by contacting us at sales@ais-digital.com . Refund requests must be submitted no later than 30 days before the end of the service period. Refunds are calculated based on actual usage up to the effective date of service termination. No refund will be issued for periods earlier than 30 days after written notice is provided by either party. Refunds are processed in accordance with AWS Marketplace billing rules.
How can we make this page better?
Legal
Vendor terms and conditions
Content disclaimer
Delivery details
64-bit (Arm) Amazon Machine Image (AMI)
Amazon Machine Image (AMI)
An AMI is a virtual image that provides the information required to launch an instance. Amazon EC2 (Elastic Compute Cloud) instances are virtual servers on which you can run your applications and workloads, offering varying combinations of CPU, memory, storage, and networking resources. You can launch as many instances from as many different AMIs as you need.
Version release notes
This release includes the production-ready AMI for deploying the LUMI platform on Amazon EC2. The platform provides a unified multi-provider AI gateway, MCP Gateway, enterprise analytics, prompt evaluation, AI skills marketplace, cost optimization, and enterprise security.
Additional details
Usage instructions
Launch an EC2 instance using the provided AMI and the recommended instance type.
The application starts automatically during instance initialization.
Before using the platform, prepare an application.properties configuration file and provide the required environment-specific settings, including PostgreSQL connection parameters, Redis configuration, encryption keys, and at least one supported AI provider.
Start the application by passing the configuration file path as the first command-line argument.
If the configuration is modified, restart the application for the changes to take effect. Model prices, feature flags, and cache policies are managed through the administration interface and do not require configuration file updates.
For a new deployment, configure the initial administrator account using the lumi.superadmins setting before the first startup.
After deployment, verify the installation using the /health endpoint and ensure that all required services are running successfully.
Support
Vendor support
For product support, customers can contact us at sales@ais-digital.com . We provide assistance with installation, configuration, and operation. All support inquiries receive a response within 24 hours during business days. Additional support options or escalation paths can be arranged upon request.
AWS infrastructure support
AWS Support is a one-on-one, fast-response support channel that is staffed 24x7x365 with experienced and technical support engineers. The service helps customers of all sizes and technical abilities to successfully utilize the products and features provided by Amazon Web Services.