3,748 tools found
An evidence-aware multi-engine math kernel - Python library and MCP server - for LLMs that need provenance-tracked mathematics, not bare numbers.
Python library that gives sub-agents a narrowed, signed credential (a Pigeon Pass) instead of a copy of your API key, and verifies it before tools run.
OKF Agent Memory is a git-native, MIT-licensed persistent memory layer for AI agents, stored as plain Markdown with BM25 search and zero API cost.
Correct small email errors while preserving voice and commitments, with optional tracking-link cleanup and an explicit change list.
Queries a compiled embodied-AI knowledge graph via REST for grounded, source-linked implementation paths - never for controlling physical robots.
Build and fix existing Laravel behavior using root-cause diagnosis, repository conventions, regression coverage, and proportionate verification.
Judgement layer over OrcaReplay's MCP server: tells an agent to read a recorded run instead of guessing what already happened.
In-browser LLM inference engine using WebGPU, with an OpenAI-compatible API, streaming, JSON mode, workers and Chrome extension support.
Assess prose for machine-generated writing signals with linguistic evidence, explicit uncertainty, and abstention on non-prose documents.
Extract document fields as structured JSON with page and confidence references, reporting absent values as not found instead of guessing.
Screen people or organizations for corroborated adverse media, PEP signals, and sanctions exposure with human-review framing.
Builds a self-contained HTML slide deck from a repo where every claim cites its source line and can be re-checked for staleness.
