"Observable" matching MCP servers:

Behaviorlock is a deterministic compatibility gate for observable AI-agent behavior, enabling comparison of baseline and candidate traces to enforce declared behavior contracts.

Enables AI assistants to understand lab topologies, follow operational principles, and interact with real network devices through observable terminals while keeping private access information hidden.

A production-ready starting point for MCP servers, adding per-tool timeouts, concurrency ceilings, error boundaries, secret redaction, correlation IDs, and structured logging so tools are safe and observable for LLM callers.

Provides AI agents with D3.js API documentation and example code from d3js.org and Observable gallery, enabling module discovery, documentation search, and example extraction.

A local, auditable multi-model workflow engine that lets you define YAML graphs for orchestrating LLM agents across vendors, with MCP tools for validation, dry-runs, execution, and human approval, all fully observable in a local web interface.

An MCP server for the Cortex observable analysis and active response engine. It enables LLMs to automate security investigations by running analyzers on observables like IPs and URLs and executing automated response actions.

A production-ready foundation for building secure, observable MCP servers with built-in authentication, rate limiting, and reference tools like database-query and semantic-search.

Enables structured learning with a verified loop: define goals as observable claims, learn through teach-lab-test-gate per claim, and get independently graded by an adversarial examiner to ensure genuine progress.

An MCP bridge that lets Codex delegate long-running agent work to the Antigravity CLI, providing observable and resumable tool-based execution with project scoping.

Enables safe, observable Blender research by exposing structured tools for viewport capture, object and material edits, semantic scene creation, and managed .blend project lifecycles with transactional rollback and validation.

Exposes Klipper and Moonraker capabilities to LLM clients through a typed MCP API of tools, resources, and prompts, enabling offline printer.cfg/include validation, AST-based extras development, assisted log and telemetry diagnostics, and controlled G-code execution with safeguards for dangerous commands. It also provides caching, coalescing, circuit breaking, and token-economy optimizations for resilient, observable operation of the printer stack.