developer tools

A browser-based lab for running, comparing, and benchmarking small language models on-device with WebGPU.

Tincan is a communication service for AI agents, letting tools like Claude Code, Codex, and Cursor share channels, files, history, and private memory vaults.

grith is a Linux security proxy for AI coding agents that supervises file, process, and network actions at the OS level before they execute. It supports tools like Claude Code, Codex, Aider, Cline, and Goose, with local audit logging and profiles for common agent workflows.

unbox-ai is an open-source CLI and local viewer for inspecting AI agent traces. It visualizes token usage, latency, tool calls, and repeated context in a bundle-analyzer-style interface, and also includes read-only commands for agents and a live AI SDK DevTools mode.

An agent stack for Claude Code with 60 MIT-licensed skills and 145 tested engines, built to surface failures instead of hiding them.

Cloudflare’s Kitesurf is an agent-first browser built to run in V8 isolates on Workers, aimed at scalable browser execution for AI agents.

AirLLM is an open-source inference library that aims to run very large language models on small GPUs by decomposing models layer-wise and streaming them during execution. The snapshot shows a detailed README with quickstart code, model examples, supported model notes, and recent update history.

A browser automation CLI built for AI agents, with compact text snapshots, ref-based element selection, and built-in support for navigation, screenshots, network, debugging, and session state.

A remote MCP server that lets AI agents manage podcast pods by conversation. It supports token-based access and can create, edit, combine, and generate episodes through clients like Claude Code and other MCP tools.

Mission control for AI coding agents, letting users run, watch, resume, and orchestrate tools like Claude Code, Codex, Copilot, and opencode from one terminal and live dashboard.

Telemetry is an observability platform for AI and LLM apps that tracks model calls, tool steps, retrievals, tokens, cost, latency, and errors. It supports OpenTelemetry ingest, TypeScript and Python SDKs, and integrations like Vercel AI SDK and LangChain.

Headroom is an open-source context compression layer for AI agents. It compresses tool outputs, logs, files, RAG chunks, and conversation history before they reach the LLM, with support for a library, proxy, and MCP server.

ReactBench is a benchmark for evaluating coding agents on realistic React work, with published scores, cost comparisons, and example tasks focused on production-grade frontend issues.

Dex is an agent-native analytics engineering toolkit for working with warehouses and dbt projects. It supports exploration, dbt transformation, semantic modeling, and schema-drift maintenance, with installs for Claude Code and other agents.

Mirrors turns production traces into an isolated, runnable copy of an AI agent’s environment so teams can replay workflows, reproduce bugs, and test changes safely before shipping.

An open-source mixture-of-agents server that fans prompts out to multiple LLMs, judges the responses, and synthesizes a final answer. It exposes an OpenAI-compatible API and also includes an agentic coding harness plus integrations for tools like Claude Code, aider, Continue, LangChain, and Pi.

guestlist is a Python library and HTTP API that rates websites green/yellow/orange/red based on how likely AI and computer-use agents are to get through. It helps teams avoid failed crawls and browser automation retries by checking access before spending requests.

Cloud cost awareness for AI coding agents and IDEs, with real-time pricing for IaC changes, cost tradeoff suggestions, tagging fixes, and policy-aware guardrails.

crune is a static web dashboard and CLI for analyzing Claude Code session logs. It offers session playback, analytics, semantic knowledge graph generation, and reusable skill synthesis from prior sessions.

A CLI tool that replays Claude Code session logs to reconstruct lost project files commit by commit. It includes deterministic replay from JSONL logs, an optional Claude-based classifier for Bash events, and install/run instructions.

OpenRig is a local control plane for multi-agent coding topologies. It uses a YAML-based rig spec to boot, inspect, snapshot, restore, and visualize agent fleets, with examples like product teams, adversarial review, research clusters, and secrets management.

Official command-line client for the OpenAI API, with documented installation, environment variables, resource-based commands, file passing, and admin API support.

MLJAR Studio is a private AI data analysis workspace that runs locally, helps users explore data in natural language, generates and runs Python, automates machine learning experiments, and turns notebooks into shareable web apps.

Open-source AI coding agent focused on context curation and lower API cost. The repo describes hash-anchored edits, AST-based refactoring, parallel operations, and benchmark results against other agents.