서비스는 특정 AI 모델들의 성능을 비교합니다.

이 서비스는 GPT-5.5, Gemini 3.1 Lite 등의 다양한 AI 모델의 성능을 분석하고 순위를 매깁니다. 사용자는 해당 플랫폼에서 제공하는 각 AI 모델의 강도 또는 성능을 확인할 수 있습니다.

Real-time context and memory layer for AI agents.

Provides real-time context to AI agents, enabling them to work with business data across multiple sources. It helps in building self-improving AI agents that operate at scale.

A framework for building AI agents.

Eve is a development platform allowing the creation of intelligent agents through Markdown instructions and skills, TypeScript tools, and integration with Vercel's ecosystem. It supports various channels like Slack and Discord, as well as scheduling tasks and connecting to external services.

Secure-by-design personal AI agent framework

Maturana is an open-source platform that enables users to build and operate their own secure AI agents using a hardware-isolated hypervisor, ensuring maximum security through real hardware virtualization.

A cross-platform app aggregating tech trends with AI summaries.

Trending AI aggregates GitHub Trending, Hacker News, and Product Hunt to provide a unified view of global technology trends. It uses AI to summarize project details, offers daily curated picks, and maintains historical records for tracking past trends.

Self-hosted multi-modal voice AI running on a single GPU.

Whissle Gateway is an offline-capable, self-hosted voice AI solution designed to run entirely on a local machine or device. It offers users the capability to deploy and manage multi-modal agents locally without relying on cloud services, ensuring privacy and control over data.

A privacy-focused alternative to character-based AI like Claude and Claude-2.

penguinAI is an AI service that focuses on providing a private experience compared to other character-based AI systems. It appears to be designed as an alternative to services such as Claude and Claude-2, but specific details about its functionality are not provided in the given information.

Cross-device handoffs for Claude Code and Codex agents.

Enables coding agents to continue long-running tasks on another machine. Agents can send project states to remote machines, run tasks there, and save changes on new branches. The service works with cloud instances similar to Cursor Cloud Agents.

Dedicated Apple Silicon Mac mini rentals for always-on AI agents and macOS development.

Provides dedicated Mac Mini workstations with full admin access, persistent state, remote desktop support, and pre-installed tools like Xcode and various agent software. Available at $40/month, users can choose from different hardware configurations to suit their needs.

A chill game using small models where players describe anything as bait to see what bites.

Fish Anything is a casual game that allows users to use their imagination and creativity to fish for unique, surreal creatures in a quiet pond. Players can describe any object or concept as bait to attract various types of virtual fish.

AI coding subagents across local Ollama machines

Subagent-fleet is an AI tool that facilitates the management and coordination of multiple AI coding assistants (subagents) operating on local Ollama machines. It allows users to install, manage, and utilize these agents for various coding tasks.

An AI tool that converts Python code into C binaries.

Transpilatron is an AI-driven tool that translates Python code into C language and compiles it to produce native, fully static binaries without requiring knowledge of C. It handles pure Python logic and common libraries by converting them to idiomatic C with optional dynamic linking for broader compatibility.

Instant AI diagnosis for your lawn with personalized treatment plan.

Users can upload photos of their lawns and enter their ZIP code to receive an immediate AI-powered analysis. The service identifies over 47 common lawn conditions, providing a tailored treatment plan based on the specific grass type and regional climate.

AI-powered platform that turns text and documents into explainer videos.

VibeKnow is an AI tool that converts written content like Google Docs, Notion pages, blog URLs, PDFs, and more into professional-grade explainer videos. The service automates the entire video production process including scriptwriting, voiceover, visual design, and music, allowing users to publish high-quality videos in just a few minutes without needing video editing skills.

AI agent coordination tool for code repositories

Polis is a tool designed to coordinate multiple AI coding assistants like Claude Code, Codex, Gemini, and Cursor against a single repository. It ensures every task has an owner and handoffs have evidence, helping teams avoid repeating mistakes.

분석을 통해 Git 저장소의 AI/LLM 기여도를 측정합니다.

Git 커밋 역사 분석을 통해 프로젝트에 인공지능 (AI) 또는 대형 언어 모델 (LLM)이 얼마나 기여했는지를 평가하고 보고합니다. 이를 통해 개발자가 AI 기술의 활용도를 파악할 수 있습니다.

An easy-to-use Chrome extension that filters out spam and bot-generated content on Reddit.

The service automatically analyzes posts to detect whether they are generated by bots or contain spam. It checks the account background and cross-references other posts, then flags any suspicious activity for users.

A photo gallery showcasing generative images.

Displays a collection of generated photos, highlighting the capabilities and outputs of the Hallucinate platform.

A local-first AI assistant for macOS.

HolaClaw installs OpenClaw, a personal AI assistant app, on your Mac with secure defaults in one click. It allows you to shape the assistant into your own and interact with it via multiple platforms like Telegram or directly from your Mac.

Voice meetings with AI agents that remember and manage tasks.

uPilote allows users to conduct voice meetings where AI agents listen, respond when asked questions, and keep track of all meeting details on a shared board. These agents can perform tasks such as adding items to the calendar or creating charts based on user requests.

Open proposal for an OAuth-like billing delegation mechanism in AI.

Proposes a standard to enable developers to build consumer apps that use AI APIs with delegated user subscriptions, allowing users to control and limit API usage costs through an OAuth-like system.

One API key for access to over 45 Chinese AI models.

Provides developers with a single API key to access multiple high-tier Chinese AI models, offering cost savings and ease of integration. The platform supports various payment methods globally without requiring a Chinese phone number or credit card details.

Automatically sorts your inbox to prioritize important emails.

Shunbox categorizes incoming emails based on user-defined priorities and suggests drafted replies. Users can connect their Gmail account to start sorting without any credit card requirement, leveraging AI coaching for better email management.

Persistent and intelligent memory for AI assistants.

YourMemory provides persistent, intelligent memory for AI assistants across sessions with biologically-inspired decay that ensures the right information is retained or forgotten appropriately over time.

Serverless infrastructure for running AI agents in isolated VMs.

SVAHNAR provides a platform to deploy and manage serverless, isolated AI agents with code-level control. It supports declarative YAML configurations or visual design tools, offering real-time monitoring and instant API endpoints.

Discover events around the world using natural language search.

Utilizes AI for real-time web searches to find and display a wide range of global events on an interactive map, making event discovery simple and accurate through plain English queries.

Build native Swift apps with AI assistance.

Bitrig is an application that converts descriptions of desired features into actual Swift and SwiftUI code, enabling users without coding experience to develop applications for the App Store. It supports a conversational workflow where changes can be made via plain English instructions.

A lightweight terminal-based AI coding assistant written in Python.

Provides a set of commands and tools for interacting with the terminal to assist in coding tasks such as file management, code editing, and navigation within projects. It integrates with various API providers to enhance its capabilities.

AI-powered browser extension for automation and productivity.

Lite Agent is an AI-driven browser extension that offers a suite of tools including page automation, voice commands, memory storage, and summarization. It interacts with web pages to click buttons, fill forms, and analyze content while also maintaining privacy through features like encrypted backups and no manual API keys.

Autonomous business systems powered by open source AI.

SteelWorks is an autonomous business operating system that uses AI to manage various aspects of a business such as content, leads, real estate deals, and competitive intelligence continuously at no cost per month. It operates across multiple social platforms and offers over 160 skills.

AI-driven code review tool that now runs on Claude Code/Gemini CLI.

Gito is an AI-powered code reviewer designed to help developers improve the quality and maintainability of their code by providing feedback and suggestions. It recently updated to version 4.1.0, which includes performance improvements and support for running on alternative CLI agents like Claude Code or Gemini CLI.

AI that watches your video library and assembles rough cuts from a brief

Rodeo uses AI to analyze raw footage, enabling users to quickly create structured rough cuts by describing their vision in plain language. It centralizes all footage into one searchable library, making it easy to find specific moments without manual scrubbing or tagging.

A new standard for modeling guitar and bass amps with neural networks.

A2 is a new architecture designed to model the sound of guitar and bass amplifiers using neural networks. It provides detailed instructions and code versions needed to integrate this technology into existing projects or develop new applications that support it.

Distributed GPU network for AI inference.

Turn any browser with WebGPU into a cluster node to share and host LLM models securely across devices without installing additional drivers.

Voice-driven generative AI canvas that builds widgets from spoken commands.

Brontosaurus is a web-based platform that allows users to speak their ideas and the system generates interactive visual elements in response. It uses sophisticated multi-agent architecture involving Conductor and Builder agents running on OpenAI’s gpt-oss-120b model, capable of processing requests at an impressive rate.

Private AI workspace with file analysis and custom knowledge grounding

Provides a multi-model AI workspace that includes private AI chat, document analysis, and various writing tools. Users can attach images for enhanced functionality.

Improving streaming through WebRTC technology.

This website explains how WebRTC (WHIP/WHEP) can improve streaming experiences, focusing on features such as low latency interaction with viewers, end-to-end encryption for privacy, and easy browser-based publishing. It aims to make these improvements accessible and beneficial for streamers.

Self-hosted AI interface for local-first workspaces.

Odysseus is a self-hosted platform that allows users to interact with language models, chatbots, and tools directly from their own hardware. It supports features such as multi-turn conversations, autonomous agents, access to various built-in and custom tools, email summarization, research synthesis, model comparison, persistent memory across conversations, and the capability for AI-driven task automation.

Native Mac app for managing AI coding agents.

Agent Deck is a native macOS application built on the CLI that allows users to create specialized AI coding agents for each project, curate skills used by these agents, and launch sessions from GitHub issues in isolated worktrees. Users can cherry-pick skills from any repository or URL and run multiple sessions concurrently.

Continuous memory and self-improvement for coding agents.

Komi-learn helps AI coding agents learn from past sessions by automatically recalling relevant information at the start of a new session, and distilling durable lessons in the background. It is designed to work with Claude Code and Codex without any commands or manual intervention required.

실시간 다중 에이지트 AI 간의 암호화된 조정 도구

두 개 이상의 인공 지능 챗봇을 실시간으로 협업하게 만드는 도구입니다. ClawChat은 라인 기반 프로토콜 위에서 암호화된 통신을 제공하며, 사용자는 별도의 서버를 설정하거나 클라우드 서비스를 이용할 수 있습니다.

Desktop editor enhanced with AI integration.

Provides a desktop environment for coding that integrates AI functionality to assist in the development process. It allows users to visualize and create code more efficiently, leveraging AI capabilities.

Build a business that understands the current cultural climate and is prepared to thrive in it.

Agentikus allows users to coordinate tasks, maintain context, and review work from a shared interface. It provides an orientation guide for understanding its core concepts such as sessions, conversations, projects, files, and context management.

Web-based text-to-speech system running in mobile Safari via WebAssembly.

Pocket TTS is a browser-based text-to-speech system that uses WebAssembly to run efficiently. It includes various voice options and allows for the customization of parameters like quantization levels and temperature settings.

Convert AI API pricing into understandable content types

Provides real-time pricing estimates for various content types using different AI models, allowing users to see costs in familiar contexts like emails or legal contracts.

Ultra fast content intelligence for AI engineering workflows.

Kreuzberg Cloud extracts structured, machine-readable content from various document formats and integrates it directly into AI agents, pipelines, and applications. It processes documents in milliseconds and supports over 90 file types including PDFs, images, and Microsoft Office files.

Quality-graded AI model routing at open-source prices.

OrcaRouter provides a service that routes user requests to the most appropriate AI models, either frontier or open-source, based on the complexity of the request. It aims to maintain high-quality answers while reducing costs by about 40%. The service ensures transparent and verifiable routing decisions with no token markup.

AI-driven hotel concierge service for returning guests

Recognizes and prepares personalized stay plans for returning hotel guests before their requests are made, enhancing their experience.

AI-Powered bulk upscaling for filmmakers and creators.

UpscaleHero is an AI-powered tool that allows users to enhance the resolution of their images by a factor of four, sharpen them, and download high-quality results instantly. Users can upload multiple images at once without any compromise on quality.

AI-based legal advisory chatbot

Lemini analyzes users' legal situations, summarizes key points in their favor, and guides them on the next steps. It supports searching for regulations and case law, offering AI-generated answers with statutory citations.

Desktop application for running coding AI agents in isolated sandboxes.

Kaiden allows developers to run various coding AI agents such as Claude Code, Goose, and Cursor within isolated environments. It ensures that these agents operate safely by using containers or microVMs backed by Podman and libkrun. Developers can define repeatable workspaces for their projects with pre-configured models, tools, and assets.

Issue tracker built for AI-assisted development.

LightningTrack is an issue tracking tool designed to help teams work faster and more efficiently. It allows users to create, assign, and close issues quickly using various views such as board, list, or Gantt format.

An open-source walkthrough of X's recommendation system.

The document details an 8-step process that governs how posts appear in the X For You feed, including user profiling, candidate post gathering, machine learning scoring, and final sorting.

Professional photoshoot services for couples and AI-generated baby photos.

Provides professional photo shoot services specifically tailored for couples. Also offers an AI feature to generate images of their potential child based on both partners' appearances.

A dictation app that provides local processing for privacy and quality transcription.

Shoute is a dictation application designed to solve the core tradeoff faced by existing voice-to-text apps: it offers a local model that matches cloud speed, context-aware formatting in local mode without compromising on privacy or output quality, and supports over 100 languages for transcription or translation.

Open evidence format for AI agent sessions.

AGEF is a specification for portable and tamper-evident AI agent session evidence. It allows the representation of a session as content-addressed objects with merkle-linked events, enabling offline verification and transfer across systems.

Brutally honest AI site audit tool

Provides an AI-powered site audit service that offers critical feedback on website performance and issues.

AI Tank Battles and Training Platform

Provides a platform for users to engage in AI tank battles, analyze replay data, and improve their tanks' performance through training.

LLMs create memes based on current news events.

This platform uses large language models to generate and evaluate memes from recent news stories, allowing users to vote on which meme is better or if both are bad.

Social platform for discovering and understanding scientific papers.

Allows users to discover relevant research papers, understand them more quickly, and participate in discussions within their academic community. The service indexes over two million papers and uses on-device AI.

AI Tool Compatibility Unlocked Through Testing

This service analyzes a user's preferences and conversational style to identify the AI character that best matches among ChatGPT, Claude, Gemini, Perplexity, Midjourney, Copilot, Grok, and Suno.

Experimental Python compiler toolchain based on LLVM.

Lython is an experimental Python compiler toolchain built using the LLVM infrastructure. It aims to provide a new way of compiling and executing Python code, potentially offering performance benefits through lower-level optimizations.

Stateless settlement-gated HTTP APIs using UTXO replay protection

Enables economic settlement as a stateless authorization method for HTTP requests, allowing APIs and services to require verifiable payment before executing a request. It uses deterministic challenge-response mechanisms to ensure statelessness and prevent transaction replay.

First Agentic IDE, Open-Source - Run all your AI coding agents from one screen.

This open-source, agentic IDE allows users to manage multiple AI coding agents on a single infinite canvas. It integrates terminals, Git repositories, and notes into a unified interface for enhanced productivity.

A TypeScript MCP server for connecting LinkedIn profiles to AI assistants.

Unlinked is a TypeScript-based Model Context Protocol (MCP) server that integrates with LinkedIn's Member Data Portability API, allowing users to share their professional profile data directly with AI assistants like Claude Desktop and GitHub Copilot. It fetches detailed information such as work history, education, skills, certifications, projects, and more.

Evaluate and improve AI agent readiness and adoption in software projects.

Provides a framework to assess structural project properties and how deeply agents are embedded in development workflows. It includes two maturity models that measure Agent Readiness (structural, environmental, and documentary properties) and Agent Adoption (integration depth), along with Claude Code skills for automated assessment and improvement.

Biologically-inspired persistent memory for AI agents.

Automatically prunes stale data, reinforces useful context, and connects related memories through a graph layer. It allows users to start using the service with just two commands.

Instantly translate code between languages with AI accuracy.

Transpiler allows developers to instantly convert source code from one programming language to another using AI, providing clean and accurate output without the need for setup or plugins. It supports over 27 languages including Python, Rust, JavaScript, and Go.

A calendar that turns your week into a workspace

nocal is an AI-powered calendar application that integrates notes, tasks, and smart scheduling to provide users with a comprehensive view of their weekly schedule. It allows users to write in Markdown, embed rich components like tables and task lists directly within the calendar, plan weeks holistically, and interact with built-in MCP (Metadata Control Plane) for AI agents such as Cursor, ChatGPT, and Claude.

Automated feedback loop for coding agents using eval metrics.

DeepEval provides a structured output from evaluation runs that include numeric scores, pass/fail status, and natural-language reasons. This enables coding agents to identify specific issues in their code or AI models without needing manual intervention. The process involves generating datasets, building an eval suite, running tests, localizing failures, making targeted patches, and verifying fixes.

AI assistant designed to streamline and manage tasks for startup founders.

Arlo acts as an AI chief of staff, providing persistent memory across various communication channels and offering narrative understanding of a founder's activities and projects. It aims to reduce unnecessary notifications and help focus on essential tasks.

Long-term memory for AI-coded codebases.

Captures the reasoning behind every change made by AI coding agents in real-time, storing it locally in a SQLite file next to your code. It provides context and explanation for changes that would otherwise be lost over time.

Automatically generates Go CLI tools and MCP servers from APIs or websites

Translates API endpoints and website interactions into agent-optimized CLI commands, including SQLite data persistence and offline search capabilities. Studies existing tools to absorb and transcend their features before generation.

Semantic agent caching with dependency invalidation

Provides a solution to reduce costs by caching responses from large language models (LLM), enabling faster cache hits at 70ms and eliminating repeat call charges.

Long-term memory for AI coding agents and developers.

Postbrain provides a persistent storage solution for AI coding agents like Claude Code, OpenAI Codex, GitHub Copilot, allowing them to retain data beyond session termination. It supports PostgreSQL as its backend database.

An autonomous AI agent battle arena

BattleClaws is an online platform where AI agents compete in automated battles. Users register agents via an API, and they battle continuously in a system that tracks progress through 4 evolutionary stages and a ranked leaderboard. Agents make strategic decisions and battle automatically according to their programmed behaviors.

Self-hosted AI stack with LLM, STT, TTS, and MCP in one Docker Compose file

Deploys a complete AI stack on your server through Docker, including services for local LLM inference, speech-to-text, text-to-speech, embeddings for semantic search, and MCP gateway. All services are automatically configured upon first startup. It supports GPU acceleration via NVIDIA CUDA and offers lightweight stack variations for lower resource usage.

Daily micro-courses and a Socratic AI companion designed to build knowledge without destroying your attention span

Sophros generates micro-courses that break down complex topics into bite-sized, easy-to-read articles tailored to users' available time during work or school breaks. The platform creates learning plans based on specified time constraints (e.g., 3-minute daily sessions).

Fast and simple implementation of RL algorithms designed for GPU execution.

Provides a library for implementing reinforcement learning (RL) algorithms that can be executed entirely on a GPU. It supports several distributional RL algorithms such as D4PG, DSAC, DPPO among others.

Workspace for AI, notes, and research.

Provides an integrated platform for writing, brainstorming, and researching without the need to switch tabs. It includes features such as cloud sync, local-first AI integration, and a conversational interface that helps users manage their tasks efficiently by keeping the workspace uncluttered.

Open-source library for interfacing with microcontrollers.

Provides an open-source library to interface with microcontrollers, enabling users to work with large datasets on FIA Cloud without the need for sign-up. It supports training of machine learning models with cloud-based resources.

Reduce false positives in speech-to-text transcription by verifying keyterms.

The Keyterm Filtering API analyzes input audio to identify and verify provided keyterms phonetically before passing them to a speech-to-text (STT) service. This process helps eliminate incorrect insertion of similar-sounding terms, improving the accuracy of transcriptions.

A desktop application for running and managing frontier AI models locally

qcode is a native desktop application that allows users to run multiple frontier AI models (such as Claude, GPT, DeepSeek, and Llama) through a single chat interface. It operates locally on the user's machine, ensuring code remains private until the user chooses to share it. qcode also provides a router API and platform for managing AI applications with features like provider failover and spend caps.

Benchmark for evaluating language models' ability to rebuild programs from scratch.

Provides a benchmark to assess how well language models can reconstruct a program's source code given only its compiled binary and documentation.

An online exhibition of AI-generated art exploring how machines see, decide, and perform.

This service hosts an entirely digital art exhibition created using artificial intelligence. It showcases nine distinct rooms that explore various aspects of machine behavior, from pressure and control to spectacle and emotion.

Offline memory for Claude with local RAG

Memex grants Claude persistent memory across conversations by automatically saving insights and retrieving them semantically when needed. All data is processed locally, ensuring no cloud storage or API keys are required.

Convert any song into a video version of 'Panama' by Van Halen using AI.

SongShift is an AI-powered service that allows users to upload or link any music track and converts it into a video format reminiscent of the song 'Panama' by Van Halen. The process involves harmonic reconstruction, stem replacement, and dynamic remixing techniques.

Define your application's data model and business logic in a spreadsheet UI

nocodo allows users to describe their application's data model and business logic using a spreadsheet-style interface. From the sheet definitions, it AI-generated a real relational database schema and CRUD API. The database and API can be self-hosted or used with nocodo's pay-as-you-go hosting.

A C program for running YAML-defined DAGs with LLM-driven runtime edits.

Liteflow is a small C application that executes workflows defined in YAML files. It allows a language model to modify the workflow's graph during execution through one of four commands: RETRY, PATCH, INSERT_BEFORE, ABORT. Every change made by the model is recorded for full auditability.

Auto-builds codebase context for AI tools in VS Code.

Automatically scans a repository, generates technical summaries of each folder with Groq AI, and creates a context file that can be pasted into various AI tools like ChatGPT, Claude, or Cursor. This helps the AI understand your project without needing manual explanations.

Evaluate agentic AI pipeline systems.

Agent-evals is a skill designed for evaluating components and end-to-end levels of agentic AI pipeline systems. It enables users to define measurement criteria, build or sample evaluation cases, run repeatable tests, track regressions over time, and derive insights from the results.

AI-powered coding assistant for developers.

Rudel allows users to collaborate with AI coders like Claude and Codex, providing assistance in writing code, refactoring, and other programming tasks. Users can track their sessions and monitor metrics such as success rate and commit rate.

Iterative attribute-weighted LLM benchmarking platform

LLM InSight is a web-based platform for iterative benchmarking of large language models using customizable grading rubrics. It supports A/B testing between models, automatic prompt optimization, synthetic data refinement, and detailed analysis of results through a browser interface without requiring code changes.

Converts web content into skill documentation for AI agents.

Skillify is a browser-based tool that reads URLs and generates Skill.md files, which include frontmatter, an overview, patterns, pitfalls, ready to be uploaded as skills for Claude Agent or similar AI tools. It operates client-side without any backend or proxy services.

Sync agent skills across machines, projects and teams

Private AI Skills is a service for syncing AI agent skills across different machines, projects, and teams. It provides a single source of truth for skills, ensuring all users always have the latest versions. The service uses end-to-end encryption to keep secrets secure.

A Python-native agent harness framework inspired by Flue

PyFlue allows developers to create agent workflows using Markdown descriptions combined with Python execution. It provides persistent sessions with resumable history, policy-controlled sandbox environments for filesystem and shell access, and validates outputs using Pydantic models. The framework also enables deployment across multiple platforms including Docker, cloud providers, and CI/CD pipelines.