"Weights & Biases" matching MCP servers:

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 open-source MCP server that enables secure AI-to-AI collaboration across organizations without sharing system prompts, private data, model weights, or internal memory.

Enables AI assistants to connect to a trading journal, analyze performance metrics, identify psychological biases, and receive continuous trading coaching.

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

Enables auditing scientific papers for methodological biases such as selection bias and p-hacking, and assessing citation credibility and research consensus.

C++26 paper reference MCP with lookup_paper, fuzzy search, and compiler_status tools over a 216-paper ISO/IEC 14882:2026 corpus. Backs a Claude Code plugin that biases the model toward final-form C++26 idioms (reflection, contracts, std::execution senders, expansion statements, #embed).

An MCP server for chemistry: it computes molecular weights and isotope distributions from chemical formulae and resolves compound name/formula lookups, backed by NIST atomic-weight data and a cached finder over multiple online and offline sources.

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.

MCP server for deep CNKI academic search, retrieving multi-dimensional data (abstracts, keywords, references, citations, similar papers, keyword weights) and enabling citation network traversal via Playwright-based API interception.

An MCP server that provides AI trading agents with persistent, outcome-weighted memory to learn from historical performance and detect behavioral biases. It enables agents to automatically adjust strategies and optimize position sizing based on context-aware recall of past trade outcomes.

Enables enhanced decision-making through hierarchical LLM analysis, using three specialized critique agents (positive, neutral, negative) that analyze proposals in parallel and synthesize them into comprehensive, actionable insights. Helps overcome single-model biases by providing multi-perspective evaluation of complex ideas and proposals.

Enables MCP clients to query Vectorworks Spotlight fixture data from automatically exported Lightwright XML files, including counts, channels, weights, and plot QC, and to write fixture patch changes back to Vectorworks.

Enables any MCP client to turn layered PSD artwork into game-ready Spine packages by inspecting and classifying layers, rigging with meshes, bones, weights, IK, physics and animation, applying procedural or After Effects-driven FX, and running runtime and mobile-budget validation with preview output. It finishes by exporting the final .spine, JSON and atlas assets so a model can drive a full art pipeline without hand-authoring rig data.

Enables reading training history and planning complete workouts on a personal calendar, including exercises with videos, sets, reps, per-set weights, supersets, notes, and block instructions. It also supports searching the exercise library, copying or moving sessions, and retrieving performance stats, personal records, and feedback.

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.

Exposes tools an assistant can call to compute four-pillar and Zi Wei Dou Shu charts, annual and monthly scores, and two-person synastry comparisons. Also back-tests its own past predictions against a user's reported life events, returning hit rate, correlation, the years it got wrong, and refitted weights.

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.

Enables an AI agent to scout local businesses, instantly re-rank cached leads under new scoring weights, explain each score variable by variable, run one-off website audits, and generate ready-to-send outreach pitches or WhatsApp openers from natural language. It exposes the same discovery/scoring engine as the CLI, so results can be listed, scored and opened in a dashboard without leaving the chat.

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.