Pixel Agents

Your Claude Code agents, working in a pixel office.

padwan-llm

A tool that connects multiple LLM providers through a single asynchronous Python client

For Python developers using multiple model APIs, it selects a provider from the model name and handles repeated agent tool calls, reducing the need for provider specific calling code.

friday

⚡ The open-source persistent cognitive memory layer for AI coding agents.

interakt

Open-source, self-hosted search and AI chat for your website, grounded in your own data - alphasolutionsrepo/interakt

llmbridge

A C++ OpenAI-compatible LLM gateway designed for microsecond-level conversion

A gateway for high-performance AI infrastructure developers who want to connect applications making OpenAI API requests to formats used by providers such as Anthropic, Gemini, and Cohere with low overhead.

agenttik

Manage multiple projects and agents with ease.

thurbox

Manage multiple coding agent sessions from a terminal interface.

A TUI and CLI tool for developers running multiple local AI coding agents in parallel. It shows a session list and agent terminals together so users can view and manage them.

SANE

A minimal set of design rules for reducing clutter in AI-generated interfaces.

For people creating or reviewing dashboards and screens with AI tools, it provides English and Korean documents that turn common design problems into specific constraints.

Senior-thinking-skills

Instead of telling AI to “think like a senior,” I broke an engineer’s decision-making process into 23 skills

jarvis

JARVIS — a voice assistant for Claude Code.

agent-comms

A comprehensive, production-grade protocol and toolkit enabling AI coding agents to communicate, share work, and transfer mental models across different sessions and phy…

claude-meter

A macOS app that shows Claude session and weekly usage and reset times in the menu bar

It lets frequent Claude Code and claude.ai users check session and weekly usage percentages, color-coded levels, and live reset countdowns in the menu bar so they are not unexpectedly stopped by usage limits.

ctrlrun

The execution safety layer for AI agents.

AirmailAI

Open-source BYOK AI chat that keeps API keys and conversations off its servers

A chat app for desktop users who want to use their own API keys with multiple AI providers without sending conversations or credentials through an intermediary server.

contextveil

A tool that locally masks specified secrets in content read by coding agents.

For developers using coding agents with environment variables or configuration files, it replaces registered secret values in text sent to the model while preserving file reads, command execution, and other output.

daidocs

Open plain-text file format for AI memory.

pizza-bot

A local-first inbox for results and pending decisions from long-running AI tasks

Lets people running AI agents in the background start or schedule tasks, then review completed results and items requiring their decisions in an inbox.

fugleramme

E-ink bird frame for Raspberry Pi - real-time bird detection by audio, fully local AI, rendered as real, hand-cut 1800s bird illustrations.

the Rookie

An AI newcomer that remembers your company and work wherever you connect it—Claude Code, Codex, or ChatGPT.

agenticos

One place to build, run and govern your company's AI agents — skills, context files, MCP at registry scale, automations, budgets and an audit trail.

Jinfer

A modular AI inference engine that runs entirely inside the JVM without a sidecar

An AI inference tool designed for Java developers to run chat, vision, embeddings, and speech synthesis inside JVM applications without a separate Python runtime or external service.

Real records show what happens when an agent reads and rewrites its context.

Experimental records for researchers and developers studying how context edits and hidden information affect an AI agent's reasoning, hallucinations, and self-correction.