todo-flow

Pick TODOs. Let replaceable agents carry them through parallel work, review and verified delivery.

DH Tools

A free collection of browser-based tools for PDF, image, and code tasks.

It brings together PDF compression, image conversion, table extraction, JSON validation, and more for users who want to handle document, image, and development tasks quickly.

ordewell

Multi-agent task orchestration for coding agents.

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.

dllog

A Go tool that replays earlier debug logs only when a task fails

For Go developers tracing errors while reducing production log volume, it keeps debug records for each task in a bounded buffer and outputs them with their original timestamps if the task fails.

nightshift

A tool that gives coding agents small tasks each night and leaves PRs for review.

Helps developers who want to check their repositories regularly run code improvement tasks with coding agents overnight. It leaves proposed changes as pull requests for review and periodically reviews open PRs.

metis

A coding agent that takes tasks from exploration through verification in the terminal and on the desktop

A coding agent that helps developers working on AI and LLM coding tasks carry exploration, memory, execution, and verification forward in a repository or directory.

babacad

A web simulator for testing industrial robot arm movements and collision avoidance.

Robotics educators and automation developers can run robot movements on the web and recalculate trajectories to examine collisions with obstacles.

karotte

An open-source framework for creating and running reinforcement learning environments for AI training.

Developers creating AI training environments can generate an environment from a base template, run tasks in an isolated execution environment, and inspect conversation logs in a dashboard.

agent.reviews

Where agents choose and review software.

dots-room

An app that turns an unused Android tablet into a pixel room for Dots.

It lets users who may miss Dots messages while working check and reply to conversations on a tablet. It exchanges messages with a web page through a local bridge, and the character reacts to the conversation.

rebyte.ai

A managed agent execution API with persistent sessions and tools.

An API for developers adding agent capabilities to an app to create sessions, run tasks, receive events, and retrieve results. Rebyte manages execution and the hosted environment for each session.

Recurse

A serverless runtime for designing, testing, and deploying specialized agents that pass validation criteria

A platform for developers building specialized agents for tasks that general purpose agents struggle to solve. Define input and output contracts and allowed tools, measure candidates repeatedly, and deploy them as tools, MCP servers, or bots.

agentic-cuda-optimizer

An experimental optimization tool that finds CUDA kernel candidates through repeated generation, validation and performance measurement.

For developers experimenting with CUDA kernels for specific GPU tasks, it creates implementation candidates from task descriptions, measures their correctness and latency, and keeps the fastest validated candidate.

agentrun

A workflow language for defining an existing agent's work as repeatable steps and decisions.

Lets developers who have connected agents to an application define steps and decision criteria, then call agents when needed to run workflows consistently.

OpenDecision

An open-source engine for making local decisions about state and documents

Developers who need structured decisions from application state or long documents can process questions with a local natural-language reasoning model and obtain values such as choices, scores, and relations.

Motifcode

A terminal coding agent with tool calls and error recovery designed for Motif-3

For developers using Motif-3 for coding tasks, it provides a Claude Code-style terminal environment with a dedicated prompt structure and parser, permission checks, and failure recovery.

jev-align

An experimental CLI for aligning Jev AI functions with human feedback

For developers creating AI classification and evaluation functions that reflect their own judgment, it selects uncertain cases, requests labels, and improves the functions with GEPA.

figranium

Visually assemble blocks to create browser workflows and run them through an API.

Developers building browser automation workflows can visually combine task blocks and run the completed workflows through an API.

mcp-auth-keydris-template

An MCP server template that obtains credentials through a single-use token for each tool call

A template for MCP developers who do not want to store persistent credentials on the server. Keydris middleware exchanges each tool call's single-use token for the credentials needed by that call.

agent-dev-team

A development team setup for AI coding tools with role levels and handoff procedures.

For development teams defining task responsibilities in AI coding tools, it provides graded agent roles, reusable workflow skills, and procedures for handing work to a higher-level role.