AI agent with streaming, tool calling, and conversation persistence. Powered by [Strands Agents SDK](https://strandsagents.com/).

A universal LLM client with transparent provider failover, streaming tool execution, pluggable reasoning, and native observability.

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

Provider-agnostic LLM client: completions, streaming, tool calls, usage and auth across API-key and OAuth providers.

Web UI for the OpenCode coding agent — chat, streaming, tool cards, and extensions over the OpenCode engine's HTTP API.

Complete TypeScript SDK for OpenRouter API with full type safety, streaming, tool calls, and advanced features

Provider-agnostic LLM client SDK — streaming, tool calling, embeddings, image generation, speech synthesis, transcription, reranking, circuit breaker, retry, reasoning normalization, cost enforcement, structured output with schema validation, type-safe to

Angular assistant panel for NewHeap AI chat endpoints with streaming, tool-call and approval cards.

Headless coding-agent runtime for unattended, host-scheduled use: zero-dependency, dual-protocol streaming, tool loop, context compaction, checkpoint/resume. The CLI is a verifier/debugger.

React UI kit for embedding Cadenya widgets: conversations, streaming, tool approvals, custom tool renderers. Built on Radix Themes.

Headless chat assistant: OpenAI-compatible LLM streaming, tool-calling, chat history, and composer helpers. **The host owns all UI.**

Styled agent-control component library for WorkerDeck: session panel, transcript with streaming, tool-call cards, permission prompts, session list, composer. Tailwind v4 + Base UI + cva. Pairs with the headless @workerdeck/react layer.

Yet Another JSON Streaming Tool

Production-grade React components for AI agent interfaces — chat, streaming, tool calls, thought traces, and status indicators.

Anthropic LLM provider for @namzu/sdk via the official @anthropic-ai/sdk. Messages API with streaming, tool-use, and extended thinking support.

TypeScript LLM client with streaming tool execution. Tools fire mid-stream. Built-in function calling works with any model—no structured outputs or native tool support required.

Repairs malformed streaming tool-call deltas from OpenAI-compatible gateways in DeepSeek Harness, before they assemble into an undispatchable call

Powerful, typed TypeScript client for the ReyGrid API with full base URL control, SSE streaming, tool calls, and pagination.

moli — a Node-first agentic CLI: streaming tool loop, sub-agents, permissions, MCP, plan mode, memory.

Provider-agnostic AI SDK for Buntok — streaming, tool calling, chat history, RAG, and middleware