Productivity

Developer Productivity & Workflow

AI-powered tools for managing context, snippets, and developer documentation.

73 tools · 73 with full reviews · 51 free or open source · Updated Sep 28, 2026

Pieces for Developers Freemium

Pieces Pieces for Developers is an AI-powered productivity platform that functions as a persistent second brain for engineers by capturing, enriching, and indexing workflow context across browsers, IDEs, and terminals. It uses a local-first engine, PiecesOS, to maintain a private, searchable knowledge base of your coding activity, allowing you to recall past solutions and project history via natural language.

agenttrace Open Source

luoyuctl AgentTrace is a local-first terminal TUI and report generator that audits the session history of various AI coding agents. It helps developers track token spend, latency, and session health while diagnosing slow runs or failed tool calls without sending sensitive data to any hosted backend.

Vibe Kanban Free

Vibe Vibe Kanban is a local-first orchestration tool that helps developers manage AI coding agents by providing a kanban board to track tasks and isolated git worktrees for execution. It allows users to run multiple agents in parallel, review diffs, and manage project workflows without the complexity of juggling multiple terminal tabs.

AI-MEMORY Open Source

AkitaOnRails AI-MEMORY gives coding agents long-term, project-scoped memory that survives across sessions and across tools. Quit Claude Code mid-task, start OpenAI Codex, Gemini CLI, OpenCode, or Cursor in the same directory, and continue without re-explaining the architecture, the failed approaches, or the open questions.

smolvm Open Source

smol machines smolvm is an embeddable, portable, branchable microVM built to run coding agents and untrusted code safely on a local machine. It boots OCI container images inside a hardware-isolated VM in under 200ms and exposes a sandboxed shell, filesystem, and networking that agents such as Claude Code, Codex, and Pi can use as their execution environment.

coop Open Source

Trail of Bits coop is a Rust CLI from Trail of Bits that runs Claude Code and Codex inside disposable, isolated virtual machines so the agents can have full tool access — Docker, git, compilers, package managers, passwordless sudo — without any risk to your host. On Linux it boots Firecracker microVMs on KVM; on macOS it uses Lima with Apple's Virtualization.framework, choosing the backend automatically. `coop up` creates or reuses an environment for the current project and syncs the workspace in, then `coop claude` or `coop codex` launches the agent with permission prompts disabled by default because the VM itself is the isolation boundary (pass --ask to restore prompts). coop forwards API keys, GitHub tokens and chosen env vars into the guest — or keeps model keys entirely host-side in proxy mode — injects CLAUDE.md, syncs ~/.codex config, supports plugins/marketplaces and MCP server configuration, understands devcontainer.json, and offers editor integration, multi-instance workflows and shell completion. A published trust model documents the VM boundary and how guest-authored data is treated as untrusted.

Claude Code Templates Open Source

Daniel Avila Claude Code Templates is a CLI and web catalog for configuring and monitoring Claude Code. It installs ready-made agents, commands, hooks, MCP integrations, skills, and settings into a project, and ships a real-time analytics dashboard, health check, and conversation monitor for Claude Code sessions.

Director Freemium

Director Director-AI is an open-core guardrail runtime designed to detect and block LLM hallucinations in high-consequence workflows. It sits between the LLM and the user, verifying outputs against governed facts using a 5-tier scoring engine before data is displayed or stored.

Claudete Freemium

Claudete Claudette is a macOS control plane application designed to orchestrate and monitor multiple Claude Code CLI instances in a single window. It allows users to manage AI agent fleets, visualize usage, and broadcast instructions across multiple sessions simultaneously without requiring extra API keys or new subscriptions.

Fractal OS Paid

Nubler Digital Fractal is an organic operating system that functions as a collaborative workspace for humans and autonomous AI agents. It provides a unified, persistent environment where agents maintain context, execute tasks across various applications, and adapt to the specific workflows of the user.

Zuse Zuse is an open-source desktop workspace designed for managing coding agents across real Git repositories. It allows users to run multiple coding tasks in parallel, maintain persistent agent sessions, and switch between various AI models while keeping context intact.

Crabtalk Cydonia 24B V4.1 is an uncensored, 24B parameter creative writing and roleplay model fine-tuned from Mistral Small 3.2. It features a 131K-token context window and is designed for long-form narrative consistency without the content refusals found in safety-tuned alternatives.

AI Product Adoption Deck Paid

aiproduct.cards The AI Product Adoption Deck is a 104-card working playbook designed to help product teams diagnose and resolve common AI adoption failures. It provides a structured, tactical workflow that includes diagnostic cards, action patterns, and workshop templates, with a companion SKILL.md file for integration into LLM-based agent workflows.

CoderPlan Freemium

CoderPlan CoderPlan is a unified AI API gateway that provides an OpenAI-compatible endpoint to access over 50 LLM models including Claude, GPT, Gemini, DeepSeek, and Grok. It is designed for developers to use with AI coding tools like Claude Code and Cursor, offering transparent, usage-based billing and detailed per-call logs.

AIFlowLearn Freemium

AIFlowLearn / 木子霖 AIFlowLearn is an AI-native platform that transforms technical documentation, articles, and GitHub repositories into structured, interactive learning collections and hands-on tasks. It features an AI-powered coach designed to guide users through engineering workflows, error troubleshooting, and project-based skill acquisition.

Faryo Open Source

Snailflyer Faryo is a lightweight project and mobile workbench designed for interacting with tmux-backed AI coding sessions, such as Codex CLI or Claude Code. It allows users to monitor agent status, send instructions, and manage approvals directly from a phone or browser without needing a full remote desktop or secondary chat interface.

Tree Ring Memory Open Source

TerminallyLazy Tree Ring Memory is a framework-agnostic, local-first memory lifecycle layer for AI agents that prevents memory from becoming raw transcript sludge. It enables agents to store, recall, and manage structured information like decisions, lessons, and durable truths using a ring-based aging model.

OpenAgentRelay Open Source

ShakespeareLabs OpenAgentRelay is a local-first tool that allows users to expose their local AI agents or automations as team-callable capabilities over a trusted LAN. It uses a small CLI to bridge communication between agents without requiring the caller to access the publisher's source code or dependencies.

Better Agent Free

Ofek Ron Better Agent is a local web-based workspace designed to centralize multiple coding agents like Claude, Codex, and Gemini into one inspectable, persistent hub. It allows developers to manage sessions, track tool calls, and orchestrate agent workflows locally while maintaining full control over data and project context.

Tura-AI Tura is a terminal-native, open-source agent runtime harness written in Rust that optimizes AI coding tasks by reducing token consumption and model round trips. It replaces standard ReAct loops with a structured macro-tool approach, allowing agents to execute complex, multi-step workflows in a single LLM turn.

fractal Open Source

Plasma AI Fractal is an open-source tool for orchestrating hierarchical, recursive agent loops that enable large-scale, long-horizon coding tasks. By organizing agents into a tree of isolated git worktrees, it allows for autonomous workflows that span multiple context windows and independently manage their own memory, lifecycle, and budget.

WolfMarkDown Open Source

WolfMark WolfMarkDown is an AI agent skill designed to transform messy AI-generated output into professional, production-ready Markdown. It uses a combination of agent-based semantic structure recovery and deterministic tooling to verify, clean, and format documents while preserving protected content like code and identifiers.

Codex Quota Overlay Open Source

cpys (maintainer; independent community project) Codex Quota Overlay is a privacy-focused desktop utility that displays remaining usage, reset times, and credits directly beside the active Codex Desktop conversation title. It functions as a lightweight, non-intrusive companion that automatically hides when Codex is inactive or minimized.

DeskCue Open Source

Alexander Kornev DeskCue is a local-first control center that provides a unified interface for reviewing and steering AI coding agents, such as Codex or Claude Code. It allows developers to monitor task progress, inspect project changes, and provide feedback from mobile or remote devices while the agents continue running on their primary workstation.

SandBase CLI Freemium

SandBase SandBase CLI is an open-source, agent-first tool and local Model Context Protocol (MCP) bridge that connects AI coding clients to over 2,000 AI models and APIs. It streamlines agent workflows by managing credentials securely and enabling model discovery, inspection, and execution directly from within supported IDEs and CLIs.

AIPM Open Source

Abhishek Srivastava AIPM is a production-grade package manager and registry designed for AI plugin primitives, including skills, agents, MCP servers, and prompts. It functions as an infrastructure layer for AI projects, offering a mental model similar to npm or Cargo with versioning, semver resolution, and dependency isolation.

Superagent Open Source

pungme Superagent is a local macOS application that provides a workspace for AI coding agents to operate within, moving beyond simple terminal interactions. It integrates a live browser, an iOS Simulator, and project file management into a unified environment where agents can execute tasks with persistent context across git worktrees.

XiuLab Inc XiuRouter is a model API gateway that allows developers to access various LLMs, including Claude, GPT, Gemini, and Grok, through a single account. It supports multiple API protocols—OpenAI Chat Completions, OpenAI Responses, Anthropic Messages, and Gemini GenerateContent—to facilitate easy integration into coding agents and applications.

QwenCloud Freemium

Alibaba Cloud QwenCloud is an AI-native platform by Alibaba Cloud that serves as a central hub for accessing Qwen and Wan model families, agent tools, and third-party models. It integrates model inference, agent development, and cloud infrastructure into a unified interface with OpenAI and Anthropic-compatible APIs for seamless developer adoption.

OrcaReplay Open Source

Continuum AI OrcaReplay is an open-source tool for recording, debugging, and forking AI agent executions by capturing agent activity at the process and socket level. It allows developers to reproduce agent runs byte-for-byte in an offline environment or switch models mid-execution to test outcomes.

Magic Patterns Freemium

Magic Patterns Magic Patterns is an AI-powered design tool that generates high-fidelity, interactive prototypes by respecting a team's existing design system, components, and brand tokens. It focuses on closing the gap between design and engineering by allowing product teams to iterate on prototypes that are ready for handoff rather than generic AI output.

Builder.io Visual Copilot Freemium

Builder.io Visual Copilot, rebranded as Fusion, is a design-to-code visual development platform that maps Figma designs to your existing component library and codebase. It enables teams to generate production-ready React, Vue, Svelte, Angular, or Qwik code that respects established design systems and coding standards.

Jackalope Freemium

Jackalope Digital LLC Jackalope is a cross-platform, open-source desktop workspace designed to orchestrate multiple AI coding agents in parallel. It centralizes project context, isolated Git worktrees, and task management to streamline workflows for developers using agents like Claude Code, Codex, and Grok.

Cabinet Freemium

Cabinet AI Cabinet is an AI-first startup operating system and knowledge base that stores all data as local Markdown files, ensuring no vendor lock-in. It allows users to build and run autonomous AI teams that have full access to their persistent knowledge base to perform research, content creation, and scheduled tasks.

PI-Desktop Open Source

vastsa PI-Desktop is a local-first, modular desktop workspace for AI coding agents that allows developers to bring their own models and work on local projects. It uses an Electron shell, a Rust host core for security and performance, and a dedicated agent sidecar to provide an inspectable, permission-gated environment for AI-assisted development.

Ruflo Open Source

ruvnet Ruflo is an AI agent meta-harness designed to orchestrate coordinated, self-learning multi-agent swarms within Claude Code and Codex environments. It acts as an execution layer providing persistent vector memory, zero-trust federation, and automated workflow governance.

Context Mode Freemium

mksglu Context Mode is an MCP server that optimizes AI agent context windows by sandboxing high-volume tool outputs, such as logs and web fetches, and persisting session memory via SQLite. It reduces context bloat by up to 98% by ensuring raw data remains in an isolated subprocess while providing the agent with relevant indexed content on demand.

Worktrunk Worktrunk is a command-line interface tool designed to simplify Git worktree management, specifically for developers running multiple AI agents in parallel. It automates common tasks like creating, switching, merging, and cleaning up worktrees while providing workflow enhancements such as hooks and CI status integration.

RTK-AI RTK is a high-performance CLI proxy that intercepts and compresses shell command output before it reaches an AI coding assistant's context window. By stripping noise and redundant boilerplate from commands like tests, git status, and builds, it helps maintain focus and efficiency in AI-assisted development sessions.

Gentleman Programming Gentle-AI is an ecosystem configurator that adapts existing AI coding agents to provide persistent memory, Specification-Driven Development (SDD), curated skills, and native bounded review capabilities. It does not install agents itself but rather enhances them to move beyond simple chat-based interactions into disciplined development workflows.

Affaan Mustafa ECC is an agent harness and operating system that structures AI coding workflows by enforcing rigorous engineering cycles. It provides a coordinated toolbox of skills, instincts, and memory management for major AI coding assistants like Claude Code, Codex, and Cursor.

Atlas Open Source

pacifio Atlas is a local-first source control platform for AI coding agents that links every git commit to the agent session that produced it. It provides shared on-device memory and a unified query interface for multi-agent workflows, allowing different agents to access common context and historical decision-making.

Archify Open Source

tt-a1i Archify is an agent skill for AI coding assistants that turns system descriptions or repository analysis into validated, interactive technical diagrams. Instead of generating raw visual code, it uses a strict, typed JSON intermediate representation to ensure that diagrams are topologically sound, verifiable, and free from hallucinations before rendering.

workmux Open Source

raine workmux is an opinionated terminal-based tool that orchestrates git worktrees and terminal multiplexers like tmux into isolated development environments. It enables developers to manage parallel AI agent workflows by providing a clean, conflict-free workspace for each task.

Spec Kit Open Source

GitHub Spec Kit is GitHub's open-source toolkit for Spec-Driven Development (SDD): instead of prompting an agent straight into code, you define what to build and why first, and the specification drives planning, task breakdown, implementation, and a final convergence check. The `specify` CLI scaffolds a project for your agent (`specify init my-project --integration copilot`), after which you invoke `/speckit-constitution`, `/speckit-specify`, `/speckit-plan`, `/speckit-tasks`, `/speckit-implement`, and `/speckit-converge` skills in the agent's chat. SDD ships in core, while bug-fixing and idea-assessment processes are bundled extensions. It works with 30+ coding agents including GitHub Copilot, Claude Code, Codex, Gemini CLI, Cursor, Windsurf, and Qwen Code, and is extensible for organization-specific processes.

Agent Skills Open Source

Addy Osmani Agent Skills is Addy Osmani's pack of production-grade engineering skills for AI coding agents. It encodes the workflows, quality gates, and judgment senior engineers apply across the development lifecycle — define, plan, build, verify, review, ship — as nine slash commands (`/spec`, `/plan`, `/build`, `/test`, `/constraints`, `/review`, `/webperf`, `/code-simplify`, `/ship`) that activate the right skills automatically. Skills such as test-driven-development, incremental-implementation, api-and-interface-design, frontend-ui-engineering, and code-review-and-quality also trigger contextually based on what the agent is doing, and each is a structured workflow with steps, verification gates, and anti-rationalization tables. `/build` can run plan-and-implement autonomously after one approval while keeping every task test-driven and individually committed. The open skills CLI installs into 70+ agents including Claude Code, Cursor, Codex, Copilot, and Cline.

9Router Open Source

decolua 9Router is a local AI gateway that sits between your coding tools and 40+ AI providers. Point Claude Code, Codex, Cursor, Cline, Copilot, Antigravity, Gemini CLI, OpenCode, or OpenClaw at `http://localhost:20128/v1` and 9Router handles format translation (OpenAI ↔ Anthropic ↔ Gemini), quota tracking, token refresh, and three-tier automatic fallback: use existing subscriptions first, drop to cheap providers such as GLM, MiniMax, or Kimi when quota runs out, then to free tiers such as Kiro, iFlow, Qwen, or OpenCode Free. It bundles RTK and Caveman-style compression of tool results to cut 20–40% of tokens, supports multi-account round-robin per provider, provider combos with sticky routing, and a web dashboard for connecting providers via OAuth or API key.

i-have-adhd Open Source

ayghri i-have-adhd is an agent skill that changes how a coding assistant writes its answers so the next action is never buried: lead with the action, number multi-step tasks, end with one concrete next step, suppress tangents, restate state every turn, give specific time estimates, make wins visible, report errors matter-of-factly, cap lists at five items, and drop preambles, recaps, and closers. It is loosely adapted from the Adult ADHD Tool Kit but framed as an output style for LLMs, not a diagnosis. Installation is documented for Claude Code (plugin marketplace), Codex (`codex plugin add`), Gemini CLI (extension or custom command), GitHub Copilot, Antigravity, Hermes, and OpenCode (server plugin plus native skills), or via `npx skills add ayghri/i-have-adhd`; invoke with `/i-have-adhd` and turn it off with 'stop adhd mode'. Forking and editing the single SKILL.md is the intended way to tune it.

Oh My Hermes Open Source

rlaope Oh My Hermes (OMH) is an operating layer that installs on top of Nous Research's Hermes Agent and keeps Hermes as the natural-language surface while adding governed workflows. `omh setup` registers 116 installable skills organized into seven capability families and nine `ulw-*` workflows for planning, research, creation, coding handoffs, operations, and project memory, each with explicit evidence boundaries so the run reports what actually happened. `omh model` (or `/omh-model` inside the Hermes TUI) sets a head model and effort per work category with editable model chains in `~/.omh/routing/model-chains.json`, and the Maestro lane dispatches coding units to Claude Code or Codex rather than hiding a coding executor behind Hermes. It works in the Hermes CLI, Hermes Desktop, and messenger integrations, ships `omh doctor` and `omh update` for maintenance, and offers an agent-installable protocol pinned to a commit SHA.

Superpowers Open Source

obra Superpowers is an agentic skills framework and software development methodology that provides structured, mandatory workflows for AI coding agents. It enforces disciplined practices like test-driven development, systematic planning, and subagent-driven development to ensure high-quality, reliable code output.

BlaBla Open Source

Petro Yakovyshyn (independent; I maintain BlaBla) BlaBla is a Rust CLI that provides executable project memory for coding agents, storing project intent as queryable contracts (behavior.bla and structure.bla) instead of relying on chat history. It verifies behavioral and structural rules and exposes a GREEN/YELLOW/RED completion gate via commands like status, explain, and finish.

OpenSpec Open Source

Fission-AI OpenSpec is a lightweight Spec-Driven Development (SDD) framework that acts as an agreement layer between you and your AI coding assistant. After `openspec init`, slash commands drive the workflow: `/opsx:explore` is a no-stakes thinking partner that reads your code and shapes a plan, `/opsx:propose` creates a per-change folder under `openspec/changes/` containing `proposal.md`, `specs/`, `design.md`, and `tasks.md` for you to review before any code is written, `/opsx:apply` implements the task checklist, and `/opsx:archive` folds the change back into the living specs in `openspec/specs/`. Specs are plain Markdown requirements (SHALL) with WHEN/THEN scenarios and no special syntax. An expanded profile adds `/opsx:new`, `/opsx:continue`, `/opsx:ff`, `/opsx:verify`, `/opsx:bulk-archive`, and `/opsx:onboard`; Stores (beta) move planning into a shared repo so cross-repo features and platform-owned requirements have one source of truth every coding agent can read. It works with 20+ assistants including Claude Code, Cursor, GitHub Copilot, Windsurf, Cline, RooCode, Codex, and Amazon Q.

BrowserSkill Open Source

Tencent BrowserSkill connects shell-capable AI coding agents (Cursor, Claude Code, Codex, OpenClaw, CodeBuddy, WorkBuddy, Pi, Hermes Agent, DeepSeek Harness) to the Chromium browser you are already logged into. It has two local pieces: the `bsk` CLI/daemon and a Chrome/Edge extension. Browser tasks run in a separate, visible Agent Window so you can keep working, agents reuse your real login state without test accounts, and touching one of your own tabs requires an explicit borrow-and-return. When a task hits a captcha, login, or confirmation dialog the agent hands control to you and resumes afterwards. `bsk install-skill` installs a harness-specific skill that teaches the agent to open pages, read text, fill forms, scrape data, run UI regression checks on a PR, and capture full-page screenshots (`bsk screenshot --full-page`). A sandboxed-agent mode keeps the daemon in a persistent host environment via a shared `BSK_HOME`.

Multica Open Source

Multica Multica is an open-source workspace where AI coding agents appear on the issue board like teammates. You assign an issue to Claude Code, Codex, Cursor, Copilot, Kimi, OpenCode or 20+ other agent CLIs; the agent runs on a daemon on your own machine or cloud box, comments progress, raises blockers and hands the diff back for human review.

Engram Freemium

Gentleman Programming Engram gives AI coding agents persistent, searchable memory across sessions. It is a single Go binary backed by SQLite + FTS5 that exposes 15 MCP tools (mem_save, mem_search, mem_context, mem_session_summary, mem_timeline...), an HTTP API, a CLI and a TUI, and works with any MCP-capable agent including Claude Code, Codex, Gemini CLI, OpenCode, Cursor, Windsurf, VS Code Copilot and Antigravity.

WebCodex Open Source

yyjeqhc WebCodex lets ChatGPT, Claude and other MCP clients work directly with repositories and developer tools that stay on your own machines. A Runner executes file, Git, command and test operations where the code lives, while the Server exposes those capabilities to the chat client, with a Console to inspect readiness and accept or reject proposed work.

shadcn lint Open Source

shadcn @shadcn/lint is an 'agent-first' linter for Tailwind v4 design systems from shadcn, the creator of shadcn/ui. You declare which components own their styling and what may change; when a coding agent breaks a rule, the error explains what is wrong and suggests the allowed fix (a size, variant, or theme token) read from your own components and theme, so the agent can correct it in one pass. It ships as an ESLint and Oxlint plugin and works with any Tailwind v4 project, shadcn/ui not required.

Codenotch Open Source

vinzdg Codenotch is a macOS app that pins a small black notch to a screen edge and draws one ring per coding assistant's usage limit — Claude Code, Cursor, Codex, Antigravity, GitHub Copilot, GLM, Grok and others — so you can see how much of a five-hour or weekly window you have burned without running /usage. Hover a ring for its limit windows and reset times; the ring's shape shows whether the agent is still working, done, or waiting on your input.

Agent-Native Open Source

Builder.io Agent-Native is Builder.io's open-source TypeScript framework for building apps that are agents and applications at the same time. Each capability is defined once as an action; the agent uses it as a tool while the UI calls it from code, and the same action is also exposed over HTTP, MCP, A2A and a CLI with shared validation, permissions and data. The runtime ships agent chat, skills, memory, jobs, observability, automations and agent teams, and the repo includes open-source example agents (Clips, Plans, Design, Content, Analytics, Chat) to start from.

Screenpipe Freemium

Screenpipe (Negentropy Labs, Inc.) Screenpipe continuously captures your screen, audio, and app activity locally, indexes it into a searchable history, and exposes it to AI agents (Claude, Codex, Cursor, etc.) via MCP, a local REST API, and markdown-based 'Pipes' automations.

antiburn Open Source

antiburn Antiburn is a free, local desktop tool that analyzes AI coding agent session logs to identify and fix excessive token usage, helping developers stay within token limits and reduce AI costs.

claude-mem Freemium

thedotmack claude-mem is a persistent-memory system for coding agents. It captures tool-usage observations during a session, compresses them with AI into semantic summaries, and injects the relevant context back into future sessions so the agent keeps continuity of project knowledge after a session ends or reconnects. Originally built for Claude Code, it now installs for OpenCode, Antigravity CLI, Cursor, OpenClaw, Grok Bot and other agents.

Oh My OpenAgent Free

Sisyphus Labs (code-yeongyu) Oh My OpenAgent (OmO, formerly oh-my-opencode) is a multi-agent orchestration layer for coding agents. The Ultimate edition plugs into OpenCode with 11 specialised agents, 54+ lifecycle hooks, built-in websearch, context7, grep_app and LSP MCPs, Team Mode and slash commands; the Light edition ports the portable pieces to Codex CLI; and OmO Native is a standalone `omo` command. Typing `ultrawork` (or `ulw`) with a prompt has the agents plan, execute and keep going until the task is done, across Claude, GPT, Kimi and other models.

CLI-Anything Open Source

HKUDS CLI-Anything turns existing software into agent-native tooling. Installed as a skill or plugin for Claude Code, Codex, Pi, OpenClaw, Cursor or Qoder, its 7-phase harness generator wraps an application — Blender, GIMP, FreeCAD, Obsidian, n8n, Godot, Zotero and dozens more — in a structured-JSON command-line interface with a generated SKILL.md, and CLI-Hub is the registry where agents and humans browse, install and update those community-built CLIs.

CC Switch Open Source

farion1231 CC Switch is a cross-platform Tauri 2 desktop app that acts as a single control surface for AI coding tools. It centrally manages the configuration files of Claude Code, Claude Desktop, Codex, Gemini CLI, Grok Build, OpenCode, OpenClaw and Hermes Agent, letting you import providers with one click, switch between 50+ built-in provider presets (official APIs, AWS Bedrock, NVIDIA NIM, community relays), and share a universal provider across apps. It also unifies MCP server and Skills management with bidirectional sync, shows usage and balance, speed-tests endpoints, offers an optional local routing layer with format conversion, hot switching, failover and request logs, and stores everything in SQLite with atomic writes.

magpie Open Source

yetone magpie is a local model picker for AI coding agents with desktop, terminal UI and CLI interfaces. It shows the models configured for Claude Code, Codex, Gemini CLI, OpenCode and other agents, changes their native configuration, and runs a local gateway that translates OpenAI and Anthropic API protocols so developers can choose providers across agents.

Hindsight Open Source

Vectorize Hindsight is an agent memory system from Vectorize built around the idea that agents should learn over time, not just recall conversation history. It exposes three operations — retain, recall and reflect — over memory banks stored in PostgreSQL with pgvector, organising knowledge as world facts, experience facts, auto-consolidated observations and user-curated mental models with entity, temporal and semantic pathways. Retrieval runs semantic, BM25, graph and temporal strategies in parallel (TEMPR). Developers add it to an existing agent with a two-line LLM-client wrapper, via Python/TypeScript SDKs or plain HTTP, an MCP server, or coding-agent integrations for Claude Code and Cursor.

Superdesign Treg, built by the Superdesign team, is an 'OpenRouter for agent tools': a tool registry and proxy that gives an AI agent one base URL and one token to call roughly 2,800 catalogued API endpoints across about 60 providers — SEO and backlinks, social and trends, people and company enrichment, ads, scraping, and image/video generation — without signing up with each vendor. Agents search the catalog by task, read the price, and call the endpoint. Teams can also register their own tools — paid API accounts, OAuth connections, vendor CLIs such as stripe, gh or vercel, and SKILL.md bundles — so every teammate's agent can use them while credentials are injected server-side and never leave the registry. The proxy relays requests rather than modelling upstream APIs, so it survives provider changes.

Impeccable Open Source

Paul Bakaus Impeccable is design guidance for AI coding agents, created by Paul Bakaus and grown out of Anthropic's frontend-design skill. Its premise is that every model trained on the same SaaS templates produces the same tells — Inter everywhere, purple-to-blue gradients, cards nested in cards, gray text on colored backgrounds — unless it is given a design language. `npx impeccable install` detects your harness folders (~/.claude, ~/.codex, ~/.grok, ~/.hermes, .cursor and more) and installs one skill plus provider-native hook manifests; `/impeccable init` then records durable product truth in PRODUCT.md so later commands know audience, purpose, constraints and voice. Twenty-four commands give you a shared vocabulary with the agent: craft, shape, document, extract, critique, audit, polish, bolder, quieter, distill, harden, onboard, animate, colorize, typeset, layout, delight, overdrive, clarify, adapt, optimize, live and generate, plus pin for shortcuts. Sixty-one deterministic detector rules, run by the CLI or a browser extension without any LLM, catch AI-generated design anti-patterns, and live mode lets you iterate on element variants directly in the browser.

Chrome DevTools MCP Open Source

Google (Chrome DevTools) Chrome DevTools MCP is the Chrome team's official Model Context Protocol server that gives coding agents the full power of Chrome DevTools against a live browser. Added to any MCP client with `npx -y chrome-devtools-mcp@latest`, it lets an agent record performance traces and extract actionable DevTools insights, analyse network requests, read console messages with source-mapped stack traces, take screenshots and drive pages reliably through Puppeteer with automatic waiting for action results. A `--slim` mode exposes only basic browser tasks, `--headless` runs without a window, and advanced options cover concurrent sessions, persistent user-data directories, attaching to an already-running Chrome over WebSocket, and debugging on Android. A standalone CLI is provided for use without MCP, and the docs include client configuration guides for Antigravity, Claude Code, Cursor and VS Code plus guidance on running it as a browser subagent. It officially supports Google Chrome and Chrome for Testing, tracking the latest Extended Stable release.

Mobile MCP Open Source

Mobile Next Mobile MCP is a Model Context Protocol server from Mobile Next that lets coding agents and LLMs drive native iOS and Android apps on simulators, emulators and real devices through one platform-agnostic interface — no XCUITest, Espresso or per-platform glue required. It is accessibility-first: agents read the native accessibility tree for fast, cheap, deterministic element access and fall back to screenshots plus coordinate taps only when needed. Tools cover taps, swipes and gestures, typing, app install/launch/terminate, deep links, orientation, hardware buttons, screen recording and structured data extraction, which makes it suitable for agent-driven UI testing, scripted flows and multi-step user journeys as well as data-entry and scraping scenarios. The docs include step-by-step install guides for Amp, Claude Code, Claude Desktop, Cline, Codex, Copilot, Cursor, Gemini CLI, Goose, Kiro, opencode, Qodo Gen, VS Code and Windsurf, and any MCP-compatible client works with the standard configuration.

OpenRig Open Source

Esoteric Labs OpenRig is a multi-agent harness that manages Claude Code and Codex sessions as one persistent system. Developers define an agent team in YAML — pods, communication edges and continuity policies — and boot it with one command; OpenRig runs the agents as ordinary tmux sessions and tracks them through a local daemon, SQLite state, a CLI, an MCP server and a dashboard. Rigs can be snapshotted, restored after reboots and changed while running.

Cua Cua is open-source infrastructure for giving AI agents computers they can use. Cua Driver lets agents inspect and operate applications on macOS, Windows and Linux through an MCP server or CLI without moving the user's cursor, Lume creates local macOS and Linux VMs on Apple Silicon, and Cua Bench creates tasks and evaluates agents. Cua Fleets provides isolated cloud desktops for running computer-use agents at scale.

FAQ

What are the best Productivity tools?

The most complete Productivity listings on ai.dosa.dev include Pieces for Developers, agenttrace, Vibe Kanban. Every tool has a detail page with pricing, key features, and a verdict - see the full list of 73 tools below.

Are Productivity tools free?

51 of the 73 Productivity tools listed here are free or open source. Others offer freemium or paid plans - check each tool's detail page for current pricing.

How is this list curated?

ai.dosa.dev is manually curated from community submissions and editorial review. Entries are enriched with pricing, features, and reviewer notes, and the directory is updated continuously.