Search

Judge every LLM response before the user sees it. AI SDK middleware and OpenAI-compatible proxy on TypeSafe's Jev.

Count tokens per section of a Markdown file, with the right tokenizer for your target platform (Claude, OpenAI, and more).

Browser speech recognition and speech synthesis (Web Speech API) for TurboWarp, with a Composition API.

Voice chat with an AI for TurboWarp: combines OpenAI Realtime, browser speech, named functions as tools, and a forced coffee break.

Real-time neural TTS for the Pi coding agent — speaks assistant responses as they stream, powered by a Rust daemon running Kokoro-82M in-process.

Sorula: make any text-to-speech sound human. A streaming layer around ElevenLabs, Cartesia, OpenAI, Deepgram, Azure, Google, Polly, PlayHT or your own audio.

TypeScript SDK for the Agriwis (蜀农) AI API — OpenAI-compatible chat completions with built-in agricultural tools.

Workspace manager for GitHub Copilot CLI sessions

直播运营每日代办分发 CLI v2：读取钉钉群聊/私聊，AI 语义筛选每日代办，结合固定模板自动创建钉钉待办、发送汇总消息，并生成云端定时（早批次+轮询）任务

Annotate webpages in Chrome and send screenshots + element context to your OpenCode session. Fork of JodusNodus/opencode-chrome-annotation.

Vue 3 image annotation component based on fabric.js — draw rect / circle / polygon annotations on images with zoom, undo/redo, echo and PNG export. 基于 fabric.js 的 Vue3 在线图片标注组件

Agent Detection & Response — tool-agnostic detection engine that intercepts coding-agent tool calls before execution

Git의 변경사항을 분석하여 commit 메시지를 자동으로 생성하는 도구

Per-delegation reasoning effort for dsh subagents: makes AgentOptions.reasoningEffort reach child requests (the agent loop silently ignored it) and adds a subagent_effort tool that accepts reasoning_effort and model per call

Background stream listeners that wake the DeepSeek Harness agent — put a watch on a command or file and receive new lines as in-session notices. Ships a daemon host so a watcher can run unattended for weeks.

Community automation-only Agent Client Protocol server for DeepSeek Harness with committed tool-activity frames (tool_call/tool_call_update) — a drop-in alternative transport to @deepseek-ai/dsh-acp

Escalation client for cascade-search: asks Jev (TypeSafe System One) through Vercel AI Gateway.

The best model for each job: one OpenRouter model ID per kind of work, with the request settings measured alongside it.

Drop-in push notifications and VoIP calling for CometChat React Native apps. JS-first: a native display engine (FCM/notifications + CallKit/PushKit) drives the OS UI; all CometChat actions run through your existing chat-sdk-react-native — no second native

Server-side AI crawler tracking and the EchoRanked visitor snippet