#runtime

Marchward is an Apache-2.0 open-source runtime authority for AI agents. It routes an agent's tool calls through one control point: credential mediation (the agent holds a single Marchward key and never your real downstream credentials), an inference spend cap enforced before the call reaches the provider, human-in-the-loop approval gates on irreversible actions, and a tamper-evident hash-chained audit log. Public repo and packages are live (npm @marchward/*, PyPI marchward, official MCP registry ai.marchward/mcp-server), with a runnable governed-LangGraph example in the repo.

Documentation and assembly review for @nhtio/adk.

Runtime Cognitive Cortex for AI coding workflows. Contorium provides workspace awareness, project memory, runtime monitoring, MCP context synchronization, and optional cognitive insights without taking control away from developers.

Noorle is a managed AI agent runtime with secure MCP gateways: built-in tools, custom WASM plugins, and 1,000+ connectors — auth, rate limits, and audit logs included.

Runtime security oracle for AI agents — scan prompts, tool calls, and outputs before execution. Pay-per-call via x402/USDC, no accounts needed.

Lightrun MCP connects AI coding assistants to live runtime context from production and staging applications for safe code-level debugging without redeploying. It lets MCP-compatible clients discover runtime sources, inspect live expression values, capture call stacks, measure execution duration, count executions, and collect numeric runtime metrics directly from the AI workflow.

Runtime policy enforcement for AI agents. Evaluate every agent action against your organization's policies before execution. Define policies once in the Hejdar dashboard, enforce everywhere via MCP. Supports observe mode (shadow/audit) and enforce mode (real-time blocking).

Persistent memory, knowledge graph, governed actions, and runtime orchestration for AI agents. 107 tools. Works locally with zero-config SQLite(no API key needed) or connects to Novyx Cloud for the full surface. Install: uvx novyx-mcp

AI agents call LLMs, invoke tools, and hit APIs — but have no built-in way to cap spend. A single agent loop can burn hundreds of dollars before anyone notices. Cycles MCP Server gives any MCP-compatible agent a runtime budget authority: tools to check, reserve, spend, and release budget before and after every costly operation. Works with Claude Desktop, Claude Code, Cursor, Windsurf, and any MCP host. Supports per-tenant budgets, soft-landing caps, and automatic heartbeat for long-running operations.

Cryptographic runtime governance for AI agents. 20 tools. Sealed policy artifacts, continuous measurement, tamper-evident proof. Ed25519 + SHA-256.

Local LLM Server with NPU Acceleration

Production-Ready MCP Server Framework • Build, deploy & scale secure AI agent infrastructure • Includes Auth, Observability, Debugger, Telemetry & Runtime • Run real-world MCPs powering AI Agents

the mcp server that run the code in Node.js container and obtain the result

Build AI Agents MCP-based

Model Context Protocol (MCP) server for container runtimes (Podman and Docker)