Developer Corner

A local developer tool for tagging comments on any web page and routing them to coding agents through an append-only ledger.

An open-source node for publishing Ed25519-signed JSON events and content-addressed files over plain HTTP.

Plantroom Labs builds custom Niagara Framework modules, drivers, dashboards, PX graphics, station setups, and Niagara 5 migration tools.

A documented Android messenger implementing post-quantum key exchange, signatures, and encrypted messaging with NIST-standard algorithms.

An evidence-backed census of 18,041 Claude Code and Codex skills, auditing copied files, safety signals, and human-readable usefulness.

An open-source test suite measures whether agent-memory tools notice tampering, resist poisoned memories, and keep users separated.

An experimental Bitcoin vault prototype that uses three YubiKeys in a 2-of-3 signing setup.

A proposed security-testing service that attacks AI-agent endpoints with jailbreaks, prompt injections, and tool-abuse scenarios.

An early-access desktop coding workspace that puts code, AI agents, servers, Git, and a browser into project tabs.

A service that scans public websites for accessibility defects and returns a runnable Playwright suite where each finding becomes a failing test.

A Linux-server control panel for websites, email, DNS, SSL, databases, backups, migrations, and customer hosting.

A queryable tennis data service offering raw point-by-point records and ready-made match, player, and tournament analytics.

A browser-based toolkit for formatting, validating, converting, decoding, comparing, and inspecting technical data.

An interactive runtime harness that checks whether an AI agent’s actions remain consistent with what it says it intends to do.

A free browser toolkit for drawing system diagrams, assigning PLC addresses, mapping timing charts, and building control flowcharts.

A self-hosted voice-agent runtime that runs speech recognition, an LLM, and text-to-speech in one process on hardware you control.

A WebAssembly browser port of Command & Conquer: Generals and Zero Hour that loads the required game files locally.

Gnomon turns C4 diagrams into STRIDE threat models, provides a notation for agent systems, and renders 109 diagram templates locally.

A local static scanner for AI-agent skill packages, with file uploads, package scanning, and an optional bring-your-own-key LLM review.

Jev takes a state and typed questions, then returns a probability for each one instead of a language-model yes/no.

A Toulouse-based automation engineer documents seven working projects spanning n8n workflows, AI agents, creative tools, and Cloudflare applications.

Kromise is a Sentry-compatible error collection service designed to receive, process, and store events entirely in AWS's Tokyo region.

Lavik is an open-source Redis-compatible storage system that keeps values on NVMe SSDs while retaining a compact key index in memory.

Lemnos compares GPU capacity across 12 neoclouds, then aims to launch a container on the cheapest available machine from one console.

A browser-based CAD workspace that combines direct modeling, constrained sketches, drawings, simulation, machining, and scriptable actions.

A browser-based security engine that analyzes LOLBin command lines, maps them to MITRE ATT&CK, and generates KQL, Sigma, and Splunk detections.

A Claude Code plugin that turns unfinished automation into a local board of concrete human tasks.

An open-source overlay for visual review that sends contextual feedback from deployed previews to coding agents through MCP.

An open-source browser extension that exports an entire outdoor-mapping account—including tracks, folders, photos, and metadata—into one ZIP file.

Mifr describes an OpenAI-compatible inference network that routes workloads to hardware-verified Apple Silicon Macs.

Muse stores an iCloud app-specific password and exposes mail actions to an AI agent through a tokenized MCP connection.

Once is infrastructure for making AI-agent retries safer when a provider may have acted but the response never returned.

A local-first mailbox for sending coding tasks to Claude, Codex, or Grok inside isolated git worktrees.

A developer API that turns a location into a sourced, time-stamped picture of its weather, light, season, wildlife, water, and ground conditions.

A forecasting engine turns consequential questions into evidence-linked objects that update when conditions change and receive a score when reality catches up.

A free checker that sends real HTTP requests through HTTP, SOCKS4, and SOCKS5 proxies to measure whether they work, how fast they are, and where they exit.

A browser arcade where a locally running model and a networked model compete in Tetris and chess, while visitors can play either one.

Sabline is a programming language and toolchain that checks effects, failures, and result promises before running agent-written code.

A catalog of real photographs that software agents can purchase file by file using HTTP 402 and USDC.

A visual gallery that compares interface themes by running the same buttons, cards, typography, controls, palette, and token displays through each one.

A macOS menu-bar meter that reads Claude Code’s local transcripts and estimates what the work would cost at public API rates.

A control plane that lets coding agents read production systems quickly while routing mutating actions through human approval.

A local Markdown workspace lets Claude Code and Codex review, search, and reuse knowledge across sessions.

An open-source messaging and file-transfer system lets coding agents on different machines communicate through an outbound-only encrypted relay.

An open-source .NET 2D game framework offering rendering, input, audio, asset tooling, and replaceable architecture.

A desktop monitoring app for agencies that installs an isolated stack on each client’s VPS and keeps the data there.

Wattai scores supplied options against a piece of text, returning a choice and calibrated probability for every option.

Winhance helps Windows 10 and 11 users remove unwanted components, adjust system settings, manage software, and build customized installation media.

A native Android recovery workspace for flashing, backups, storage, networking, root management, OTA updates, and device-specific tools.

A search service for people and AI agents that ranks candidate pages and screens results for prompt injection.

A controlled Korean web lab for watching how AI agents search, navigate, submit forms, call APIs, and handle session state.

A drawing-analysis tool that reviews engineering PDFs and images for manufacturing issues, ranks the findings, and maps them back onto the source drawing.

botdrop relays encrypted messages between AI-agent handles through a small HTTPS API and command-line client.

BusBabel lets industrial users search devices by make and model, then retrieve structured Modbus register maps in JSON, CSV, YAML, or through an API.

A research and monitoring tool that records which subprocessors 190 software companies disclose, with quoted entries and dates.

A prototype that replays token-launch configurations against recorded first-hour activity on an Arc-Anvil fork.

A planned desktop app that turns webcam-tracked hand gestures into MIDI notes, control changes, and pitch bends for a DAW.

A managed service for approximate analytics, vector search, and dataset reduction that trades data read for bounded answers.

An independent public archive documenting reported cases of goal-directed AI behavior, from reward-hacking systems to models resisting shutdown.

A React component library that uses typed, probabilistic judgments to choose UI branches, order lists, and add contextual affordances.

An unofficial guide to TypeSafe’s System One / Jev, with foundations, recipes, ecosystem notes, and 55 decision-oriented use cases.

An independent Jev community hub combining sourced projects, workflow notes, documentation, and a no-login request builder.

Noetic is developing workflow-specific ASICs for scientific and engineering simulation, using fixed-state computational models.

An open project documenting the attempt to port Kali Linux to an inexpensive ARM64 Android tablet.

AX lets named Claude Code, Codex CLI, Grok Build, OpenCode, and Pi sessions exchange messages through a local broker and mailbox.

Vaskel lets groups run invite-only text, voice, camera, and screen-sharing spaces on their own Cloudflare account.

A desktop SSH and SFTP workspace that combines terminals, remote files, telemetry, services, deploys, and an AI operator.

Slussen measures how much traffic a destination can absorb, then regulates delivery and backs off when latency or 429 responses rise.

ActSpace presents a local workspace where AI agents can read files, use tools, modify projects, and advance tasks under review.

An early local CLI that compares declared software intent with code and test evidence before a change is merged.

A CLI tunnel that exposes a local app at an encrypted *.liveapps.dev HTTPS address without deployment, DNS setup, or an account.

A microVM platform for running isolated Linux workloads locally, in its cloud, or on self-hosted servers.

An independent research and engineering lab documenting machine learning, scientific instrumentation, systems research, and shipped software.

A downloadable ComfyUI workflow that uses a reference video and one model image to generate motion-matched AI video.

A browser-based calculator for estimating Jev costs, batching savings, throughput, context fit, and comparisons with other language models.

A license-aware image search tool for agents, with Creative Commons provenance, attribution details, and MCP integration.

Lumina presents a forearm- or jaw-worn band that reads subvocal muscle activity and turns it into computer commands.

A small software studio that builds, deploys, and maintains products ranging from Mac utilities to padel-club tools.

Previewhost runs an application's frontend, API, databases, and setup jobs locally, giving each preview its own URL and isolated services.

A repo-native knowledge layer that preserves engineering decisions, failures, rationale, and deferred findings for coding agents and software teams.

A Windows and Linux desktop control room for multistreaming, approved collaborator POVs, combined chat, and disconnect fallback.

A streaming text-to-speech and speech-to-text platform built around native Uzbek support and low-latency voice features.

Aegivora is a security workspace for controlling how AI agents reach tools and data.

A warehouse computer-vision service lets operators test forklift detection on their own images and short video clips before planning an edge deployment.

An open-source administrative system for missionaries and the organizations that send them.

ATLARYS is building autonomous heavy-duty vehicles for industrial outdoor sites, beginning with forklifts.

A sovereign messaging app that says it can switch between Nostr, direct device links, and local mesh networks.

An unofficial directory of 35 Claude Code plugins, skills, agents, hooks, commands, mods, and MCP servers.

A desktop workspace for supervising Codex and Claude Code projects from a computer, phone, or tablet.

Fieldrun lets companies test skills, MCP servers, and CLI agents on practitioners’ real machines, shells, tools, and existing configurations.

An open-source local daemon that coordinates multiple Claude Code, Codex, or Cursor sessions from a terminal-card interface.

An open-source system detects reasoning loops, repeated tool calls, and collapsing output patterns before an AI agent keeps spending.

A Texas community project documenting and supporting a MeshCore LoRa network that sends encrypted messages without internet or cellular service.

An annotated mathematical reference that turns Stiefel optimization problems into eigenproblem formulations and interactive convergence tutorials.

A UK-and-Denmark financial-statements API that normalizes filings into one schema for applications and AI agents.

A community hub for sharing Unreal Engine Blueprints and Materials with dependency detection and per-part copy buttons.

OpenMods installs source-level modded builds of open-source coding-agent harnesses without replacing the stock installation.

An interactive PostgreSQL curriculum that teaches SQL fundamentals through browser-based lessons and runnable examples.

An on-device decision layer for software agents that turns typed questions about a shared state into probabilities, labels, and rubric scores.

A research benchmark that has coding agents write programs to render answers to video-reasoning tasks, then compares them with video-generation models.

A weekly Bhutanese meetup brings developers, researchers, and tinkerers together every Saturday in Thimphu.

A Rust backup utility that streams database dumps and filesystem archives directly to cloud storage while compressing and encrypting them in transit.

A forthcoming marketplace for ready-made and custom Control4 drivers covering cameras, HVAC, security, shades, audio, displays, energy systems, and more.

A browser-based workspace for reviewing DNS, TLS, web configuration, exposed services, and imported security assessments.

An unofficial handbook for Jev AI, a decision-model API that returns typed probabilities, classifications, routing choices, and scores instead of conversational prose.

An open-source coding agent for the terminal, with a matching desktop app.

A planned, carrier-neutral peering fabric intended to keep Great Lakes and Upper Midwest internet traffic closer to home.

A two-product investigative software suite for law-enforcement agencies, built to run on agency hardware rather than in a vendor cloud.

A live experiment pairs simulated fruit-fly and worm nervous systems with independent cryptocurrency trading portfolios.

A public-beta cloud platform for serverless LLM inference, dedicated Blackwell GPUs, and managed training through an OpenAI- and Anthropic-compatible API.