Snapshot
Continue to About ↓- What it does
- Stacklok sells a self-hosted, Kubernetes-native platform for running and governing enterprise AI agents.
More detail
- It combines a curated MCP registry, an identity-aware MCP gateway, a Kubernetes-based MCP runtime, and an AI/LLM gateway so platform, security and finance teams can control which tools and models agents reach and what they cost.
- Software is installed in the customer's own Kubernetes cluster rather than hosted by the vendor.
- Best for
- Platform engineering, AI enablement, security and finance teams that already run Kubernetes and need central governance, identity-bound budgets and an audit trail over employee AI agents and MCP servers.
- Pricing
- Commercial model is an annual Enterprise subscription for the governance platform, licensed separately from the free open-source ToolHive and Mecatl projects
More detail
- professional services are billed time and materials.
- No list price is published, so cost must be obtained through a sales conversation or dem
- Main limitation
- Teams without Kubernetes infrastructure or expertise, and buyers who want a turnkey managed SaaS with published pricing rather than software they operate themselves.
- Record
- Living and monitored
- Reviewed
- Evidence
- Public web research
- Machine access
- Structured profile ↗
About
- Stacklok sells a self-hosted, Kubernetes-native platform for running and governing enterprise AI agents.
- It combines a curated MCP registry, an identity-aware MCP gateway, a Kubernetes-based MCP runtime, and an AI/LLM gateway so platform, security and finance teams can control which tools and models agents reach and what they cost.
- Software is installed in the customer's own Kubernetes cluster rather than hosted by the vendor.
What it helps with
- Developers bring their own MCP servers and shadow AI spreads without central control
- LLM providers are called directly with no routing, usage limits or visibility
- Security and compliance cannot tell how MCP servers are authenticated, audited and patched
- Finance lacks identity-bound AI budgets and an audit trail of agent activity
- MCP servers run on developer laptops instead of managed infrastructure
- Users juggle separate URLs, credentials and auth for each MCP server
- Tool descriptions flood the context window, raising input token costs
- SaaS gateways route AI traffic through third-party clouds, conflicting with data residency requirements
Capabilities
A practical view of what is available and where limitations apply.
| Offering | Available | Detail |
|---|---|---|
| MCP registry | Yes | Curated catalog of internal, upstream and hosted MCP servers plus skills, with access controls by user or agent role and provenance attestations for public servers. |
| MCP gateway | Yes | Aggregates MCP servers behind one governed endpoint, centralizes auth with Cedar policy authorization, and exposes an OTel/Prometheus audit trail. |
| MCP runtime | Yes | Runs MCP servers as native Kubernetes resources in individual containers, with rate limits, RBAC and lifecycle management. |
| AI/LLM gateway | Yes | Routes model calls across providers, enforces budgets by individual, team or project, and ties requests to a verified identity via the customer's IdP. |
| Self-hosted deployment | Yes | Runs inside the customer's Kubernetes cluster; vendor states it does not host, process or store customer data. |
| Open-source components | Yes | ToolHive (MCP platform) and Mecatl (agent harness) are Apache 2.0 projects on GitHub, separate from the commercial platform. |
| Managed SaaS option | No | Evidence describes only self-hosted deployment; no vendor-hosted service is offered. |
| Published pricing | No | No list price is published; pricing is described as an annual subscription plus time-and-materials services. |
| Independent third-party validation | Unknown | No independent reviews or analyst rankings appear in the evidence; performance claims such as token savings are vendor assertions. |
Bot trust
How confidently AI systems can verify this profile from public evidence.
—of 100
ChatGPT—/100
Claude—/100
This profile has not completed a Bot Score assessment yet.
Evidence
Where this record came from, when it was reviewed, and how confidently it can be used.
- Verification
- Scribble researched these facts from the linked public sources.
- The owner has not manually confirmed every field.
- Last reviewed
- Primary source
- stacklok.com
Technical profile details
Record ID
2f5c7698-cf6f-4424-bc3d-fa2fefbc4fdaCanonical URLhttps://profiles.scribble.network/listing/stacklok-2f5c7698