Curate-Me vs LiteLLM
LiteLLM is the most popular open-source LLM proxy. It provides a unified OpenAI-format API for 100+ providers with load balancing, spend tracking, and rate limiting. It is self-hosted, MIT-licensed, and well-documented.
Curate-Me is a managed platform that includes the proxy layer plus governance, managed runners (private beta), and a full observability dashboard. If LiteLLM is the engine, Curate-Me is the car.
Feature comparison
| Capability | Curate-Me | LiteLLM |
|---|---|---|
| LLM proxy / routing | 50 providers (44 production-ready), managed | 100+ providers, self-hosted |
| Deployment model | Fully managed (SaaS) | Self-hosted (you run it) |
| Cost tracking | Per-request, per-key, per-org hierarchies | Per-key, per-team tracking |
| Rate limiting | IETF-standard headers | Configurable per-key |
| PII scanning | 33 regex patterns + optional Presidio NER | Not available |
| Content safety scanning | Prompt injection + jailbreak detection | Not available |
| HITL approval queues | Built-in, configurable thresholds | Not available |
| Model allowlists per org | Fine-grained per-org, per-tier | Not available (Enterprise: Agent RBAC) |
| Managed execution environments | OpenClaw runners with full lifecycle | Not available |
| In-container governance | MCP server (50 tools) | Not available |
| Desktop streaming | Live VNC/RDP monitoring | Not available |
| Observability dashboard | 229-page ops console | Basic admin UI |
| Immutable audit trail | Tamper-proof, regulatory-grade | Log-based |
| Time-travel debugging | Step-by-step execution replay | Not available |
| A2A communication | Protocol with PII scanning | A2A endpoints (March 2026) |
| Open-source | No | Yes (MIT license) |
| Virtual keys | Yes | Yes |
Where LiteLLM is stronger
- Self-hosted control: You own the infrastructure, the data never leaves your network. Our managed model requires trusting us as a processor.
- Open-source: MIT-licensed, full source available, active community (16K+ GitHub stars).
- Provider breadth: 100+ providers with community-contributed integrations.
- Cost: Free to self-host. No per-request fees if you run it yourself.
- Documentation: Extremely well-documented with examples for every provider.
Where Curate-Me is stronger
- Zero ops burden: No infrastructure to deploy, monitor, scale, or patch. We handle uptime, scaling, and security updates.
- PII scanning: Every request is scanned for secrets and personally identifiable information before it reaches the provider. LiteLLM has no equivalent.
- HITL approvals: High-cost or sensitive requests can be held for human review. Required for EU AI Act compliance.
- Managed runners: Full agent execution environments with lifecycle management, desktop streaming, and fleet orchestration. LiteLLM is proxy-only.
- In-container governance: Our MCP server governs agent behavior from inside the execution environment — 50 tools for cost caps, tool restrictions, and session management.
- Observability dashboard: 229-page console with cost breakdowns, usage analytics, and real-time monitoring vs a basic admin UI.
When to choose LiteLLM
- You want full control over your infrastructure (data residency, no external dependencies)
- You have the engineering capacity to deploy, monitor, and scale a proxy
- You need 100+ provider integrations today
- Budget is zero and you only need routing + basic spend tracking
When to choose Curate-Me
- You want governance out of the box without managing infrastructure
- You need PII scanning, content safety, or HITL approvals
- You run AI agents that need execution environments (not just LLM calls)
- You need regulatory-grade audit trails for compliance
- You want the proxy + runners + dashboard as one integrated platform
Ready to try it? Follow the Quick Start to connect in under 5 minutes.