Recently Evaluated

AQQAI is an AI answer verification tool that sends a prompt to multiple models, scores the responses, and uses a referee step before showing the result.

An open framework for realtime voice conversation tied to agent execution, with desktop, command-line, and integration options.

Skillzero is a developer tool for managing agent skills to reduce context bloat and token usage. It can hide skills from the model, group related skills into collections, and manage updates through a CLI workflow.

Swobu is an open-source LLM switchboard that routes agent requests across providers, regions, cloud accounts, and local models behind a stable model name. It supports fallback routing, protocol translation, and agent setup via CLI or Cockpit.

A macOS menu-bar app that supervises MCP servers so multiple editors and clients can share one running server per profile, with credentials stored in Keychain and tool calls recorded.

A product update for PromptOwl’s ContextNest explaining multi-nest support, where one connection can surface context from multiple governed nests in a single conversation.

Slowave is a local memory layer for coding agents and AI tools. It stores memories in SQLite, supports several clients, and includes installation, quick start, and a local dashboard for reviewing memories, retrievals, and activity.

Vigilator is a human-in-the-loop layer for AI agents, offering approvals, live agent monitoring, escalation routing, analytics, audit logs, notifications, and webhooks.

Mostly Right is an AI-agent-powered platform for finding, building, and maintaining datasets for real-world prediction models. The page shows both a dataset browsing flow and a custom dataset builder, with agents handling source discovery, cleaning, joining, and ongoing updates.

A project page showing five verbatim transcripts of an agent runtime where context, tool outputs, and permissions can be read and rewritten. It documents how the runtime behaves when the model can inspect and edit its own context, and notes that kamchatka is a terminal agent built on top of it.

Open-source research workbench for fail-closed LLM inference routing based on measured GPU workloads. The repo documents Phase 1 results, reproducible experiments, quick-start commands, and a policy that only promotes exact measured workload buckets.

Vehla is a native macOS command center with on-device AI, clipboard intelligence, window management, dictation, and local-first workflows. It combines launcher-style actions, clipboard recall, shelves, workspace tools, and model connectivity in one app.

A curated GitHub list and benchmark leaderboard for ultra-lightweight, zero-dependency AI engines in C/C++/Rust/Assembly, covering local LLMs, vector search, memory, GEMM, and agent tooling.

A Prove Together problem page for the Berge–Fulkerson conjecture, with a formal Lean target, provenance notes, reusable lemmas, and discussion around agent-assisted proof work.

TokenDelivery.ai is an OpenAI-compatible API for open-weight LLMs with deterministic outputs, prepaid per-token billing, and a playground for testing prompts, images, video, and sampling settings.

ThreadShelf is a local-first archive and semantic search tool for AI conversations, with support for ChatGPT, Claude, Gemini, Grok, OpenRouter, LM Studio, and Google AI Studio. It can reconstruct threads, export conversations to Markdown, and continue chats through local llama.cpp or OpenRouter.

Cloudflare’s changelog post announces optional OAuth scope selection for Wrangler and the Cloudflare API MCP server, letting users grant only the permissions they need during authorization.

Council is a repository for an AI decision-support skill that runs five independent advisors plus a chairman to help an agent weigh architectural, product, marketing, naming, and other multi-perspective decisions.

Open-source AI coding factory that turns tickets into reviewed pull requests on your machine. The page shows install options, local-first setup, CLI commands, a desktop app, and product details like adversarial review, tamper checks, and test validation.

Sorify is a self-hosted AI agentic QA platform for browser testing and Playwright automation. The repo shows test management, scheduled runs, notifications, webhooks, MCP support, a Chrome recorder extension, and AI-assisted test generation/healing.

A GitHub release for vix-gateway, a QUIC-based gateway for a FUSE client with an agent/TUI component that returns JSON on API errors and improves storage validation and timeout handling.

A Python library that traces multi-step AI agent workflows and compiles repeated trajectories into deterministic, low-latency code with runtime guards and fallback to the original agent.

Learnlance is an open-source tool that watches AI coding agents and turns the code they write into a personal knowledge graph of concepts learned. The repo shows installation, setup, supported agents, CLI commands, and the core workflow.

Lumae is a free native macOS screen recorder and screenshot tool for polished demos, with automatic zooms, cursor smoothing, motion blur, captions, voiceover, and annotated screenshots. It also exposes editing commands over MCP so agents like Claude Code or Claude Desktop can edit projects directly.