"Weights & Biases" matching MCP servers:

Enables querying W\&B Weave traces and creating W\&B Reports with text and visualizations through natural language.

Enables LLM agents to query Weights & Biases experiments, including listing projects, runs, metrics, plotting metrics, and retrieving run details.

Enables running ML experiments from a local laptop by coordinating GitHub for code, Kaggle for data and GPU training, Weights & Biases for tracking runs, and Google Drive for small artifacts.

An MCP server that enables LLMs to detect cognitive biases and logical fallacies in text, list supported biases, and provide links to relevant psychology tests.

Enables auditing the AI models and SDKs in a repository, drafting evaluation examples, ranking candidates by cost per passing answer, and generating cheapest-first routing configurations, all through read-only local tools that never make network calls. Also estimates monthly token spend and local weights/KV-cache memory so users can pick and size the cheapest model that passes their own tests.

Checks whether an open-weight LLM fits your GPU: VRAM for weights, KV cache and overhead at each quantisation, a tokens-per-second ceiling, the longest context that fits, and what would work instead when it doesn't. Reads any Hugging Face repo's config.json. Free, read-only, no API key.

Enables comparing and recommending AI models based on scenario-specific criteria such as capability, context, relative cost, and self-hosting, with explanations of the scoring weights.

Enables zero-dependency INT8 symmetric quantization of LLM weights and activations on a per-tensor or per-channel basis, reporting SNR and MSE reconstruction telemetry along with TTFT/TPOT latency and jitter metrics. Also exposes edge inference primitives such as PagedAttention block allocation, radix prefix caching, and speculative decoding verification through the Model Context Protocol.

Enables agents to perform symmetric INT8 per-tensor and per-channel quantization of LLM weights and activations, returning SNR and MSE reconstruction metrics for edge inference optimization.

A local stdio MCP server that gives assistants soft, customizable rhyme tools: comparing word and phrase rhymes, searching dictionary words by phonetics, ranking and generating English lyric candidates with contextual model scores, and reporting pinned model status. Phonetic tools work without a language model, while generation and contextual ranking optionally use pinned Qwen weights under user-defined rhyme profiles.

Tweet scoring using X's actual algorithm weights - 22 signals, 18 filters, 11 tools

MCP server for scanning AI systems and code for demographic, occupational, and geographic biases. Enables CI integration and AI agent-driven bias detection.