| 10/13 | 10 |
Claude Code skill compressing agent output into telegraphic, technical phrasing to cut output tokens about 65% while preserving accuracy.
|
| 10/14 | 9 |
Headroom compresses tool outputs, logs, RAG chunks, files, and conversation history before they reach an LLM to reduce tokens while preserving answers.
|
| 10/13 | 9 |
Persistent memory compression system that preserves and summarizes context across Claude Code sessions for continuity.
|
| 10/13 | 9 |
Tool that generates workflow, sequence, data-flow, and lifecycle architecture diagrams with light/dark themes and multi-format export.
|
| 10/12 | 8 |
Local AI app builder focused on privacy, speed, and user control with support for custom API keys.
|
| 10/10 | 8 |
CLI tool for customizing Claude Code prompts, themes, toolsets, UI, and behavior across native and npm installations.
|
| 10/10 | 8 |
Multi-agent job search system that scans listings, evaluates fit, produces structured A–H reports and scores, and tailors CVs while tracking applications locally.
|
| 10/09 | 8 |
Zero-dependency MCP server that provides safe, token-efficient read-only database access across PostgreSQL, MySQL, SQL Server, MariaDB, and SQLite.
|
| 10/08 | 8 |
Framework for building agents and LLM-powered applications with modular components and integrations.
|
| 10/08 | 8 |
Local cross-platform Claude-compatible client with TUI and desktop apps, multi-agent tooling and Telegram/Discord integrations.
|
| 10/08 | 8 |
A Ruby library providing a unified API for interacting with various AI models and services.
|
| 10/14 | 7 |
Command-line tool for interacting with the Claude Developer Platform API, sending messages and managing resources.
|
| 10/14 | 7 |
Local Anthropic API proxy routing requests to Claude Max via the official Claude Code SDK.
|
| 10/13 | 7 |
Agent harness that orchestrates diverse models for automated coding, reasoning, and task execution.
|
| 10/13 | 7 |
Zero-dependency TypeScript runtime for production AI agents, providing durable execution, persistent memory, tool calling, approvals, and resumable runs across Node, Bun, Deno, and edge.
|
| 10/13 | 7 |
Claude Code plugin showing context usage, active tools, running agents, and todo progress.
|
| 10/12 | 7 |
Coi runs AI coding tools inside an isolated Incus Linux system container with defense mechanisms that protect the host from credentials and dangerous actions.
|
| 10/12 | 7 |
Desktop dictation application converting speech to text with local and cloud model support and privacy-first design.
|
| 10/11 | 7 |
Independent desktop client for OpenCode 2, providing a focused interface for projects, sessions, requests, and local connection management on Linux and macOS.
|
| 10/11 | 7 |
Desktop client supporting multiple LLM providers, available on Windows, Mac, and Linux.
|
| 10/11 | 7 |
Run Claude Desktop's Cowork local-agent mode natively on Linux without macOS or a VM, using reverse-engineered stubs for macOS native modules.
|
| 10/09 | 7 |
Mission-Control style dashboard and workspace that monitors AI coding agent activity in real time, with cost tracking, diffs, terminal, and source-control controls.
|
| 10/09 | 7 |
Local-first AI resume workspace with job management, fact ledger to prevent hallucinations, interview prep, and multi-format resume exports built on FastAPI, React, and SQLite.
|
| 10/08 | 7 |
Go-based API server integrating Claude, Codex, Gemini and OpenAI models with Gin, DB support and Docker deployment.
|
| 10/14 | 6 |
Highly configurable AI virtual executive persona with eight specialist agents, RAG from knowledge and company documents, episodic memory, and workflow scheduling across common chat and email tools.
|
| 10/14 | 6 |
Local LLM router exposing one endpoint to proxy Anthropic and OpenAI-compatible providers, enabling tools to use Claude subscriptions and other backends.
|
| 10/13 | 6 |
OpenAI-compatible privacy proxy that masks personal data and secrets before they reach external LLMs.
|
| 10/13 | 6 |
Type-safe SDK for building provider-agnostic, multimodal AI applications with chat, streaming, agent loops, and isomorphic tools.
|
| 10/11 | 6 |
Multi-agent operating system that runs autonomous AI teams to execute, coordinate, and self-improve software projects.
|
| 10/11 | 6 |
Reactive CLI generating commit messages for Git, YADM, and Jujutsu using Ollama, ChatGPT, Gemini, Claude, Mistral, and other AI providers.
|
| 10/10 | 6 |
Containerized AI coding workstation combining Claude Code, a browser web UI, headless Playwright automation, and bundled AI CLIs plus common development tools.
|
| 10/10 | 6 |
Utility that monitors provider reset times and triggers Claude Code and Codex window rollovers immediately after 5-hour limit resets.
|
| 10/10 | 6 |
Cross-platform desktop tools suite with AI chat and provider switching plus clipboard, HTTP debugging, encoding/decoding, encryption, and forensics features.
|
| 10/08 | 6 |
MCP server that lets AI assistants manage SSH connections, run commands, transfer files, handle backups, databases, and monitor server health.
|
| 10/14 | 5 |
Persistent cloud-backed shared memory for Claude Code, OpenClaw, and Codex agents.
|
| 10/14 | 5 |
Feature-rich, dependency-free single-file Node.js statusline for Claude Code that shows model, live git state, context usage, prompt-cache status, and subscription rate limits.
|
| 10/14 | 5 |
Private, self-hosted AI coding agent with a feature-rich toolkit across CLI, web, desktop, IM, VS Code, Obsidian, SDK, and mobile interfaces.
|
| 10/14 | 5 |
AI-powered desktop writing studio for Chinese web-novel authors, offering an agentic creative engine with long-range plot memory for planning, drafting, and managing serials.
|
| 10/14 | 5 |
Desktop mission control that runs multiple isolated Claude Code and Codex sessions in parallel with saved configs, per-session tracking, and cost and attention visibility.
|
| 10/13 | 5 |
Local-first desktop tray app that routes coding clients through a token-saving optimization pipeline and shows savings analytics, primarily for Claude Code.
|
| 10/12 | 5 |
Multi-account Claude Code proxy that rotates Claude Max or API key accounts automatically based on session and weekly quota limits.
|
| 10/12 | 5 |
Local gateway that connects AI clients once, records every request with cost and routing, redacts secrets, blocks dangerous tool calls, and enables safe upstream switching.
|
| 10/11 | 5 |
Smart LLM router that locally routes requests to the cheapest capable model across 30+ providers, reducing inference costs.
|
| 10/11 | 5 |
Desktop AI agent for Windows, macOS, and Linux that generates interactive drawings, reads PDFs, and includes built-in browser, video, and file-command tools.
|
| 10/11 | 5 |
Claude Code and other harness agent workflows ported from pstack, routing tasks to verified playbooks that keep code concise, simple, and improved.
|
| 10/11 | 5 |
A cross-device tmux session manager that provides a unified TUI dashboard to start, resume, and supervise AI coding agent sessions across devices.
|
| 10/09 | 5 |
Unofficial repackaging of Anthropic’s Claude Desktop for Linux distros, adding Computer Use, custom themes, multiple profiles, quick entry, and a Linux hardware feature.
|
| 10/14 | 4 |
macOS tool that runs multiple isolated Claude Desktop and Claude Code accounts side by side with separate logins, chats, settings, and connectors.
|
| 10/08 | 4 |
Modern AI gateway providing unified OpenAI, Anthropic, and AI SDK-compatible APIs with tracing, project management, and developer tools.
|
| 10/11 | 3 |
LoongPort provides an API key and base_url configuration helper for cc-switch relay providers, using a local SQLite store and a Tauri-based CLI.
|
| 10/10 | 3 |
Terminal-native workspace that runs multiple coding tasks in parallel using isolated git worktrees and persistent agent or shell sessions.
|
| 10/10 | 3 |
Claude Code/Codex/Cursor welfare-style router list with free quota, model and pricing auto-updates plus one-click configuration scripts.
|
| 10/13 | 2 |
Local-first workspace that connects multiple AI models and teammate tool approvals, keeping conversations, run evidence, and results together by project channel.
|