We check every product on this page once a day and record whether it still answers.
An open-core platform for testing, simulating and monitoring multi-step AI agents in production, including an open-source framework for running adversarial conversation simulations before launch.
Open source
What to know▼
How it works
An open-core platform to test, simulate and monitor AI agents end-to-end, including multi-turn, multi-step, multi-agent systems, with an open-source framework (Scenario) for simulating full multi-turn agent conversations — including adversarial and edge-case paths — before production.
What's different
Tests agent behavior across full interactions the way software teams test software, rather than just grading single LLM outputs like most competitors.
Best for
Teams building complex or multi-agent AI systems who need to catch broken behavior before users do.
We check every product on this page once a day and record whether it still answers.
A free, open-source tool that applies custom visual skins to Codex Desktop without modifying the app's official bundle.
Watch out · macOS and Windows.
FreeOpen source
What to know▼
How it works
Applies immersive backgrounds and visual styling to Codex Desktop through a local CDP session, without modifying the official app bundle or code signature; restores stock appearance with one click.
We check every product on this page once a day and record whether it still answers.
Syncs your existing coding-agent subscriptions across web, desktop, terminal and mobile so you can start a session on one device and continue on another.
FreeOpen source
What to know▼
How it works
Chats sync across all surfaces; runs on web, desktop, terminal, and mobile.
Pricing
Free forever, open source.
Best for
Developers who already pay for multiple coding-agent subscriptions and switch devices often.
We check every product on this page once a day and record whether it still answers.
A CLI that actually runs your test suite before letting an AI coding agent's merge-ready claim through, catching agents that say tests passed when they didn't.
Open source
What to know▼
How it works
A CLI that reads an AI coding agent's claim that tests passed, actually runs your real checks (pytest, npm test, go test, etc.), and blocks the merge if the agent's story doesn't match the command output.
What's different
The verdict comes from real command output, not a regex match on the agent's claimed text.
Best for
Developers using AI coding agents who want a guardrail against merge-ready claims that were never actually tested.
We check every product on this page once a day and record whether it still answers.
A local terminal dashboard that manages multiple Claude Code agent sessions in one view, with worktree isolation and diff review, instead of juggling separate windows.
Works offlineOpen source
What to know▼
How it works
Runs locally and open source; install with a single npx command, with all sessions shown on one live terminal wall.
Best for
Developers running several Claude Code agents at once.
We check every product on this page once a day and record whether it still answers.
Web-based terminal multiplexer for monitoring and controlling multiple AI coding agent sessions, including on mobile.
Open source
What to know▼
How it works
A web terminal multiplexer for monitoring and interacting with multiple AI agent sessions (OpenCode, Claude Code, Codex, Copilot, etc.), including a Kanban board, file browser, push notifications, and on-demand git worktrees.
We check every product on this page once a day and record whether it still answers.
A drop-in PyTorch layer for continual learning that lets similar tasks share sub-networks for transfer while keeping dissimilar tasks isolated, avoiding both forgetting and the need for a replay buffer.
Open source
What to know▼
How it works
Learns task similarity from context so similar tasks share sub-networks while dissimilar ones stay isolated, with no replay buffer.
What's different
Reproducible benchmarks disclose where the method works and where it doesn't.
Pricing
Apache-2.0 license (open source).
Best for
PyTorch users needing continual learning without catastrophic forgetting.
We check every product on this page once a day and record whether it still answers.
Open-source terminal supervisor that manages planning and worker AI agents for Codex, Claude Code, and Ollama sessions.
Open source
What to know▼
How it works
An open-source terminal supervisor where a planning agent proposes versioned plans you approve, then isolated worker agents implement the work while it handles scheduling, review, and recovery.
We check every product on this page once a day and record whether it still answers.
WordPress/WooCommerce plugin that lets you manage your whole site in plain English through a chat terminal in the dashboard — orders, products, posts, plugins, themes, users, SEO and database, via 96 tools.
FreeOpen source
What to know▼
How it works
Type commands like 'show me today's orders' or 'create a product at $29.99' in a chat terminal inside WP admin; bring your own Anthropic API key.
We check every product on this page once a day and record whether it still answers.
A debugger for AI agent runs that shows exactly what context each model call saw, for developers troubleshooting agents built with Claude Code, Codex, or Cursor.
Works offlineOpen source
What to know▼
How it works
Captures every step of an agent run, compares two runs, and can fork from any step to test changes
What's different
Local-first and open source, so traces never leave your machine
We check every product on this page once a day and record whether it still answers.
A floating always-on-top Windows window that runs Claude Code with screen awareness, so developers can ask about what's on screen and have it edit files or run commands.
Open source
What to know▼
How it works
A frameless, always-on-top Windows chat window for Claude Code, summoned by hotkey, that captures and reads your monitors before answering, then can edit files and run commands using your own Claude Code subscription.
What's different
Runs the full Claude Code agent on your existing subscription rather than requiring a separate API key, and has screen awareness rather than just chat.
Pricing
Open source (MIT).
Best for
Developers on Windows who want a floating Claude Code assistant that can see their screen while they work.
We check every product on this page once a day and record whether it still answers.
An open-source terminal coding agent with full file, command and app access on your machine; you supply your own Claude, GPT or other API key to power it.
Watch out · Grants full machine access — file reads, command execution, and app control — governed by granular permissions rather than a sandbox.
Open source
What to know▼
How it works
A terminal-native coding agent with full machine access — reads files, runs commands, edits code, opens apps, and searches the web — with granular permissions on each action, persistent memory, and voice control.
What's different
Runs natively in the terminal with no cloud IDE or web sandbox in between.
Pricing
Free to start with open models included (Deepseek, GLM, Minimax, Kimi); bring your own API key for Claude, GPT, or other providers.
Best for
Developers who want a terminal-based agent with direct, broad access to their own machine.
Watch out
Grants full machine access — file reads, command execution, and app control — governed by granular permissions rather than a sandbox.
We check every product on this page once a day and record whether it still answers.
A free, open-source local app that lets multiple AI coding agents (Claude Code, Cursor, Codex, etc.) collaborate on one codebase over MCP while the user watches and approves each step.
FreeWorks offlineOpen source
What to know▼
How it works
Runs locally with no API keys or cloud account required; coordinates coding agents over MCP with step-by-step user approval.
Pricing
Free and open source.
Best for
Developers using multiple AI coding agents on the same codebase who want oversight of each step.
We check every product on this page once a day and record whether it still answers.
An open-source, self-hosted tool that lets two developers pair-program together on the same live Claude Code session, with the collaboration backend running in their own Cloudflare account.
Open sourceSelf-hosted
What to know▼
How it works
A fork of Mosaic/cmux adding self-hosted real-time pair programming; two developers share one live Claude Code session, with the backend deployed to their own Cloudflare account.
Best for
Development teams who want real-time pair programming with Claude Code without relying on a third-party hosted backend.
We check every product on this page once a day and record whether it still answers.
An open-source bridge that lets Claude Code tasks run from the Cowork chat app with approval gates and cost caps.
Open source
What to know▼
How it works
Lets Claude Code tasks run from the Cowork chat app, with approval gates, cost caps, permission scoping, interactive debugging, and live status, requiring no separate infrastructure or databases.
Pricing
Open source.
Best for
Teams using Claude Code and Cowork who want safety guardrails without infrastructure overhead.
We check every product on this page once a day and record whether it still answers.
Three-pane merge conflict resolver for VS Code/Cursor with optional AI explanation, free and open source.
FreeOpen source
What to know▼
How it works
A three-pane merge resolver (yours | editable result | theirs) for VS Code and Cursor, with scroll-synced word diffs, authorship chips, a history timeline, and one-click apply for non-conflicting changes; optional AI reads your repo to explain and resolve conflicts.
What's different
Nothing hits git until you apply.
Pricing
Free and open source.
Best for
Developers resolving merge conflicts in VS Code or Cursor.
We check every product on this page once a day and record whether it still answers.
Turns a Figma frame link into baseline Swift UIKit code for AI coding agents like Cursor, Claude or Codex to build from. An open-source developer tool for iOS engineers, not a mobile app itself.
Open source
What to know▼
How it works
Give an AI coding agent a Figma frame link and it builds a clean cached design package, then generates baseline Swift UIKit code for Cursor, Claude, or Codex, with support for custom color/font/icon maps, base classes, full-screen coverage gates, and buffering against Figma 429 rate limits.
What's different
Works from a Figma link instead of screenshots or a raw file dump; open source, so you own the wiring for production APIs.
Best for
iOS engineers using AI coding agents to turn Figma designs into UIKit code.
We check every product on this page once a day and record whether it still answers.
An open-source, browser-based video editor built so AI agents like Claude can edit a JSON timeline over MCP while you direct — a technical tool for developers building agent-driven video workflows.
Nothing to installOpen source
What to know▼
How it works
An open-source, browser-based, Premiere-style video editor where AI agents like Claude edit by patching a JSON timeline over MCP, with the UI live-reloading in about 150ms.
What's different
Runs in the browser with no install or build step, and is built specifically so an AI agent can drive edits while the human directs.
Best for
Developers building agent-driven video editing workflows.
We check every product on this page once a day and record whether it still answers.
An open-source local control plane and MCP server for Cursor that recalls past fixes and playbooks, runs coding loops until tests pass, and can be driven from Telegram or integrated with tools like Jira and PostHog.
Open source
What to know▼
How it works
Ingests chats from Cursor, Claude Code, and Kiro, runs hybrid recall over past fixes, and runs coding loops until tests pass, exposed as a single MCP server with optional integrations.
We check every product on this page once a day and record whether it still answers.
A free, open-source tool for managing tasks across multiple AI coding agent terminals instead of juggling tabs.
FreeOpen source
What to know▼
How it works
Organizes tasks across multiple AI coding agent terminals so you focus on the task instead of juggling 20-40 terminal tabs across worktrees, working with any AI subscription you already have.
We check every product on this page once a day and record whether it still answers.
An open-source approval layer that intercepts AI agents before payments, refunds, emails or account changes so a human can approve, block or escalate the action first.
Open source
What to know▼
How it works
An open-source control layer that intercepts AI agent actions — payments, refunds, emails, account changes and other irreversible actions — and lets a human approve, block or escalate before execution.
What's different
Open-source and focused specifically on gating irreversible actions rather than general agent monitoring.
Best for
Developers building AI agents who need a human-in-the-loop checkpoint before high-stakes actions execute.
We check every product on this page once a day and record whether it still answers.
A coordination layer that hands out leases on parts of a codebase to multiple AI coding agents (Claude Code, Codex, Cursor) so they don't overwrite each other's work.
Open sourceSelf-hosted
What to know▼
How it works
Shared MCP hub gives agents leases on code areas plus a live presence feed
What's different
Open source and self-hostable
Best for
Teams running multiple AI coding agents on the same repo
We check every product on this page once a day and record whether it still answers.
An open-source approval inbox that pauses AI agent actions until a human approves, edits or rejects them via Slack, Discord, Telegram or the web - self-hostable for teams running agents.
Open sourceSelf-hosted
What to know▼
How it works
An approval inbox that pauses AI agent actions, such as sending an email, posting a comment, or running a command, until a human approves, edits or rejects them via Slack, Discord, Telegram, mobile or web.
What's different
The approval gate is a data dependency the agent can't skip, not just a prompt instruction.
Pricing
Open source (MIT), self-hostable.
Best for
Teams running AI agents who want a human-in-the-loop approval step before actions execute.
We check every product on this page once a day and record whether it still answers.
A self-hosted, open-source widget you embed in your app with one script tag, letting users request changes by text, voice or photo, which a Claude Code session then edits and commits. Built for developers who want an in-app AI coding agent they control.
Open sourceSelf-hosted
What to know▼
How it works
A single script tag embeds a prompt bar in your app; a user's typed, dictated, or photographed request triggers a Claude Code session on your own server that edits your codebase and commits the change.
What's different
Open source, self-hosted, and uses your own API keys.
Best for
Developers who want an in-app AI coding agent they fully control.
We check every product on this page once a day and record whether it still answers.
An open-source, local-first CLI that records which lines of code were written by AI, by which model and tool, building a deterministic attribution ledger.
Works offlineOpen source
What to know▼
How it works
Records which lines of code were AI-written, which model and tool via which prompt, giving one ledger across Cursor, Claude Code, and other coding agents.
What's different
Deterministic and local-first; agents can query the CLI to surface context hidden behind existing code.
We check every product on this page once a day and record whether it still answers.
A free, open-source, git-tracked memory layer that carries context between AI coding tools like Claude Code and Cursor.
FreeOpen source
What to know▼
How it works
A git-tracked memory layer that carries context between AI coding tools like Claude Code, Cursor, and Antigravity, so context compounds across tools instead of resetting each time you switch.
What's different
Portable across multiple AI coding tools via git tracking rather than being locked into one tool's memory.
Pricing
Free and MIT open source.
Best for
Developers who switch between multiple AI coding tools and want shared context to carry over.
We check every product on this page once a day and record whether it still answers.
An open-source framework that turns your existing AI coding agent into a persistent virtual cofounder that remembers your venture across sessions, for solo technical founders.
Open source
What to know▼
How it works
An open-source install turns an existing AI coding agent — Claude Code, Cursor, Antigravity, Gemini CLI, GitHub Copilot — into a persistent virtual cofounder: one line installs it, `shotgun-init` injects it into a folder or codebase, and after a one-time onboarding it remembers your venture, builds your product, and keeps your data organized.
What's different
Open-source and works on top of whatever AI agent you already use, rather than being its own separate assistant.
Best for
Solo technical founders building without a human cofounder.
We check every product on this page once a day and record whether it still answers.
An open-source, Git-native IDE for running and reviewing AI coding agents — shows what changed in the logic, not just the diff. Built for developers working with AI agents.
Open source
What to know▼
How it works
A Git-native IDE for running AI coding agents: it tracks the agent's changes at the function and class level and compares the resulting code against your original intent, rather than just showing a line diff.
What's different
Where Git shows lines changed, Aura is built to show what changed in the logic — letting you verify a task was actually completed before committing.
Pricing
Open source.
Best for
Developers who run AI coding agents and want to verify logic-level correctness before trusting a commit.
We check every product on this page once a day and record whether it still answers.
A macOS menu-bar tracker for Claude and Codex AI usage limits that warns you before you hit the cap. Built for developers using Claude Code or Codex CLI.
Open sourceNo tracking
What to know▼
How it works
A macOS menu-bar app showing live session/weekly/monthly usage windows for Claude and OpenAI Codex, with reset countdowns, a 90% redline warning, and a forecast of when you'll run out at your current pace. It reads the CLI logins you already have.
What's different
One-click switch between providers, 9 themes, 5 layouts, self-updating via GitHub releases, native Swift, essentially one source file, zero telemetry.
Pricing
Open source, MIT licensed.
Best for
Developers using Claude Code or Codex CLI who want to avoid hitting a surprise usage limit mid-task.
We check every product on this page once a day and record whether it still answers.
An open-source AI coding agent that works in your editor, terminal, or via Discord and Telegram, with approvals and diff review before it touches code. For developers who want to supervise an agent, not just chat with one.
Open source
What to know▼
How it works
An open-source, local-first AI coding agent for VS Code-compatible editors, CLI/TUI, and messengers via Ether Live Mode (Discord, Telegram), keeping a reviewable task timeline with approvals, checkpoints and pending diffs, and routing across BYOK or Grik-hosted models.
What's different
Keeps a reviewable approval timeline instead of running as an editor-only assistant, with swarm workers for recurring checks and remote follow-ups.
Best for
Developers who want to supervise and approve an AI agent's changes rather than just chat with it.
We check every product on this page once a day and record whether it still answers.
An open-source local MCP server that gives AI coding assistants like Claude Code, Cursor and Copilot a live map of your AWS setup, database schemas and infrastructure-as-code.
Watch out · Read-only and local-only by design, but it does connect to your live AWS account to build its map.
Works offlineOpen source
What to know▼
How it works
An open-source MCP server that gives AI coding assistants like Claude Code, Cursor, and Copilot a live, read-only map of your AWS services, database schemas, and infrastructure-as-code, including exact partition keys, Lambda event shapes, and misconfiguration warnings.
What's different
One command, zero config, runs 100% locally with no data leaving your machine.
Pricing
Open source.
Best for
Developers using AI coding assistants who want those assistants to understand their real AWS infrastructure instead of guessing.
Watch out
Read-only and local-only by design, but it does connect to your live AWS account to build its map.
We check every product on this page once a day and record whether it still answers.
A Python SDK for scripting multi-agent coding workflows, letting different AI coding agents implement, review, and compete on tasks in isolated git worktrees.
Open source
What to know▼
How it works
Lets developers script the full development process in Python so agents like Claude Code and Codex implement independently, review one another, revise, run neutral tests, compete, and safely apply the winner, each working in an isolated Git worktree.
What's different
Open source, includes 40 multi-agent LLM paper workflows.
We check every product on this page once a day and record whether it still answers.
Open-source control plane that routes coding tasks across AI agents like Claude Code, Cursor, and Codex with cost tracking.
Open source
What to know▼
How it works
An open-source control plane that routes coding tasks across AI coding agents (Claude Code, Cursor, Codex, Wrangler, etc.), runs multi-agent workflows, caches to reduce repeated token spend, falls back on billing/availability issues, and shows cost per task with files touched.
What's different
Gives one measurable layer above multiple separate agent tools instead of running each independently.
We check every product on this page once a day and record whether it still answers.
A hardware-isolated VM for running Claude Code, Codex, or Grok Build so your API keys and AWS credentials stay protected from malware, exfiltration, and prompt injection.
FreeOpen source
What to know▼
How it works
Runs agentic coding tools in an isolated VM on Apple Silicon or via OpenAI/Anthropic, reachable by app, CLI, or SSH.
Pricing
Free and open source.
Best for
Developers running AI coding agents who want their credentials isolated from supply-chain and prompt-injection attacks.
We check every product on this page once a day and record whether it still answers.
A free, open-source macOS menu-bar app that translates whatever you're typing into natural, professional language and pastes it back at your cursor.
FreeNo sign-upOpen source
What to know▼
How it works
Type in your own language, press a shortcut, and the fluent translation is auto-pasted; use your own API key (Claude, GPT, Gemini) or run fully local via LM Studio.
Pricing
Free and open source, no subscription.
Best for
People writing in a non-native language across any macOS app.
We check every product on this page once a day and record whether it still answers.
A self-hosted, open-source AI coding agent that runs on your local Ollama models and can be driven from Slack, Telegram, Discord, or the web, with human-in-the-loop plan approval before it edits code.
FreeData stays with youOpen source
What to know▼
How it works
Uses a ReAct architecture with human-in-the-loop plan approval, deterministic edits, and PR creation; supports DeepSeek, Llama, Qwen, Mistral, OpenAI, Claude, and Gemini models.
We check every product on this page once a day and record whether it still answers.
Gives a repo one place to define its own rules instead of scattered AGENTS.md and prompt files, plus a CLI command to check they're followed; an open-source tool for developers using AI coding agents.
Open source
What to know▼
How it works
Gives a repo one place to define its own rules — instead of rules scattered across AGENTS.md, prompt.md, and other markdown files — and one CLI command to check the code follows them.
What's different
Open source; doesn't replace your coding agent, just gives the repo a single source of truth for its rules.
Best for
Developers using AI coding agents whose repo rules have drifted across multiple markdown files.
We check every product on this page once a day and record whether it still answers.
Turns a $30 ESP32 touchscreen into a physical control surface for answering Claude Code's prompts, for developers running multiple coding-agent sessions, instead of tabbing back to a terminal to respond.
FreeOpen source
What to know▼
How it works
Tap the touchscreen to answer Claude Code's prompts, run macros, or talk to it across every session running.
Pricing
Free, open-source, personal project, not a company.
We check every product on this page once a day and record whether it still answers.
An open-source API for programmatically posting to Bluesky, X, Pinterest (and soon Meta and LinkedIn) in one call, with a native MCP server for Claude and Cursor. Built for developers and AI agents automating social posting.
Open source
What to know▼
How it works
A single API call publishes a post to Bluesky, X, and Pinterest, with Meta and LinkedIn in review; it runs preflight checks against roughly 80 platform-specific rules before posting and includes a native MCP server for Claude, Cursor, and Claude Code.
What's different
Open source (Apache 2.0), with pinned API versions, default idempotency, and transparent error envelopes.
Best for
Developers and AI agents that need to automate posting to multiple social platforms from code.
We check every product on this page once a day and record whether it still answers.
An open-source dashboard and automation layer for Claude Code that tracks repos and sessions, enforces standing checks, and blocks repeat mistakes — built for developers already using Claude Code.
Open source
What to know▼
How it works
An open-source dashboard for Claude Code that tracks every repo, session, and routine, holds standing checks (like "tests always green") true by fixing drift automatically, runs a manager per repo, blocks repeat mistakes machine-wide, and can rescue crashed sessions. Installed with one command, npx claude-basecamp.
What's different
Fully local with zero dependencies and released to the public domain, and only escalates real decisions rather than every drift.
Pricing
Open-source and public domain.
Best for
Developers running multiple Claude Code sessions and repos who want automated drift-fixing and cross-repo oversight.
We check every product on this page once a day and record whether it still answers.
A terminal tool to browse, search, resume and convert AI coding sessions across Claude Code, Codex, Gemini CLI and OpenCode -- local-first, open source, with no telemetry.
Works offlineOpen sourceNo tracking
What to know▼
How it works
A terminal picker to browse, search, resume and branch AI coding sessions across Claude Code, Codex, Gemini CLI and OpenCode, including converting a conversation from one agent's session format to another's.
What's different
Runs 100% locally as a single static Go binary with no telemetry.
Pricing
Free and open source (MIT license).
Best for
Developers juggling multiple AI coding tools who want one place to manage sessions.
We check every product on this page once a day and record whether it still answers.
An open-source Claude Code skill that enforces spec-first, evidence-based AI coding workflows.
Open source
What to know▼
How it works
A Claude Code skill that forces spec-first AI development with evidence required at every step, addressing 14 tracked ways AI coding agents cut corners.
Pricing
Free, open-source, zero dependencies.
Best for
Developers using Claude Code who want to prevent AI shortcuts in their workflow.
