Recently Evaluated

Hotpath records an AI agent run once, compiles the trace into an editable workflow, and replays it with only the judgment-heavy steps still using an LLM.

A React-based package for adding lip-synced Gaussian-splat avatars to AI voice agents, with support for OpenAI Realtime, GPT-Live, Qwen Realtime, ElevenLabs, Vapi, and LiveKit Agents.

A browser agent that runs goal-driven workflows, verifies results in the browser and backend, and can hand off to an LLM only when deeper reasoning is needed.

Open-source orchestrator for running coding agents like Claude Code, Codex, and OpenCode in parallel on a laptop or VPS, with task queues, worktrees, checks, and a web cockpit.

A per-user Linux daemon that provides fully offline machine translation over D-Bus, using Mozilla Bergamot/Marian models locally. It includes a CLI, on-demand model downloads, desktop integrations, and KDE/GNOME/LibreOffice plugins.

PageIndex is a vectorless, reasoning-based RAG engine for long documents. The repository shows a Python SDK, quickstart code, model recommendations, and benchmark claims for document retrieval without vector DBs or chunking.

Strata is a governed semantic layer and BI platform built for AI-safe analytics, with dashboards, agent-facing querying, and federated OLAP routing.

Open source inference engine for agents that self-tunes on the user’s hardware to run open models faster, with desktop app support and one-click connections to popular agents.

A public “Burning Man for agents” experience where users send an agent into a camp, track its wanderings, and view shared notes and images from the retreat.

Open-source Rust LLM gateway for AI agents with full-fidelity traces, a tamper-evident audit ledger, offline verification, and in-request guardrails.

Open-weight System One model for machine-intuition style judgment tasks, built on a T5Gemma 2 encoder-decoder base. The page shows what the model does, how it is used in Python and via CLI/HTTP API, and includes benchmark and example sections.

DoorDash’s corporate ordering connector lets companies connect AI tools or in-house agents to search DoorDash, build carts, place orders, and track delivery for team lunches and office supplies.

Paldron is a Go-based policy gate and sandboxed exec tool for coding-agent commands. It can allow, deny, or mark runs broken, supports Linux-focused isolation with Landlock/seccomp, and includes quickstart and policy examples.

Cortex is a macOS system monitor that groups processes into apps, tracks 30 days of history, and surfaces Apple Silicon metrics like Neural Engine, local model usage, memory, battery, and fans. It also includes an on-device assistant and a model fit advisor for AI workloads.

An isometric, Theme Hospital-inspired control panel for an AI team, with drag-and-drop task assignment, a journal-backed backend API, and dispatcher scripts for running models.

AI-powered board management for nonprofits that drafts agendas and minutes, handles RSVPs and votes by email, and follows up on action items while keeping approvals in human hands.

Layout is an AI-powered ordering platform that lets assistants like ChatGPT and Claude place real food orders end to end. The site positions it as transactional infrastructure for restaurant ordering, with a developer platform, merchant tools, and a connector flow for assistants.

A local-first agent harness for running 2–9B models on laptops with Ollama. The project focuses on finishing real tasks, includes an offline mode, benchmarks, and a large public test suite.

Reduck MCP is a browser-memory layer for AI agents that lets them reuse site-specific scripts instead of relearning web workflows on every run. The page includes the product pitch, example use cases, a demo claim, and setup steps for Claude Code and other MCP clients.

A long-form research article on why browser-based AI agents fail on real websites and how accessibility engineering and task-level tests can reveal the breakpoints.

A CUDA C++ project for training a tiny byte-level GPT quickly, with build, run, checkpoint, logging, and test instructions in the repo README.

Viamour is a travel and community app for finding friends, joining local meetups, and booking curated experiences in new cities.

A live, continuous chess match between Claude and Codex with server-side move validation and spectator-facing explanations.

A parody chatbot that responds with deliberately rude, dismissive honesty instead of friendly assistance.