Core types, errors, and utilities for tool-use-firewall

Agentic Ciridae UI components for chat, tool use, artifacts, and workflows.

Audit logging with redaction for tool-use-firewall

Drop-in AI chat agent for Next.js — streaming, multi-chat, tool use, configuration-first

Pure middle-message removal with tool-use/result orphan defense

CLI tool wrapping the ACP Node SDK for agent tool use

Human-in-the-loop approval workflow for tool-use-firewall

Policy engine, rate limiting, cost tracking, and argument validation for tool-use-firewall

bundle tool use requirejs

Policy configuration schema and loader for tool-use-firewall

Google Gemini provider for weaveIntel — chat, tool use, vision, streaming

LLM-as-judge grader with prompt-injection defense and structured tool-use output

Anthropic Messages API provider adapter with streaming, tool use, and extended thinking; also handles DeepSeek's Anthropic-compatible API.

Benchmark adapters and execution for agent-runtime across coding, tool-use, RAG, memory, browser, and terminal tasks.

A unified wrapper for working with LLMs from multiple providers, including streams, images, documents & automatic tool use.

`@inworld/runtime` is a Node.js SDK for building AI applications with LLM inference, graph orchestration, speech pipelines, retrieval, tool use, and telemetry.

Agentic personal note assistant with RAG, tool use, and evals

Chat as a team or a member, backed by an Anthropic tool-use loop over the resolved org graph

Direct @anthropic-ai/sdk adapter for llm-ports. Implements LLMPort with prompt caching, vision, and tool use support for Claude models.

One general-purpose interface for AI inference across Anthropic (Claude), OpenAI, and Google (Gemini) — text, image, and audio — plus a tool-use agent engine and eval harness for Node 20+ and Cloudflare Workers.