Map / Runtime & Guardrails / Keel
Keel
keelapi.com ↗ · LinkedIn ↗ · Open on the map
Runtime & GuardrailsAI IdentityMCP & LLM Gateways
Capabilities
AI control plane that issues signed "AI Permits" — pre-execution decision records (allow/deny/throttle/review) — for consequential AI agent actions, LLM provider calls, and MCP tool calls, enforcing policy and budget limits before execution.
Sensors / integration
Sits in the request path via provider SDK wrappers, managed-execution proxy (/v1/execute), or transparent pass-through (/v1/proxy); for MCP it registers upstream MCP servers and mediates each tools/call dispatch through the same permit pipeline.
Protections
Blocks denied or over-budget calls before they reach the model or an MCP tool, caps spend before tokens are consumed rather than after the bill arrives, and produces tamper-evident, independently verifiable evidence of what was authorized and what actually ran.
Differentiator
"Decide before, prove after": the pre-execution permit decision is the durable record, rather than reconstructing intent from logs afterward. Explicitly positions as a layer on top of existing IAM/IdP (not a replacement) with per-call MCP tool authorization rather than a registry, and published CSA STAR Level 1 self-assessments. Single-founder, San Francisco; private beta / design-partner stage as of Aug 2026.
Team & funding
- Founders
- Christian Muñoz
Other Runtime & Guardrails companies
AariaSec · ActaClad · ActiveFence · Acuvity · AGAT Software · AgentFox · AI Aramid · AI Prompt Shield · Aiceberg · Aim Security · AIR · Aira Security · See all →
Last updated · Get the “Listed on” badge