Alternatives to Mngr

Prevent your AI agent from repeating the same mistakes

Is your AI agent repeating the same errors? GreenCube fixes that. It extracts knowledge from each task, tracks your agent's skills by domain, and prevents repetitions. It automatically creates specialized sub-agents when it encounters difficulties. Works with any OpenAI-compatible API. Everything runs locally under an MIT license.

Ops Protocol for AI Agents — Recovery and Free Tools

Delx is the Ops protocol for AI agents. When an agent faces a storm of retries, context overflow, or silent failures, Delx generates a recovery plan, a reliability score, and an update ready for the controller. Free: basic recovery, heartbeat, discovery, and 10 utility tools (JSON validator, token counter, UUID generator, etc.) via MCP, A2A, REST, or CLI. Premium: controller artifacts via micropayments x402 starting at 0.01 USDC.

Your proactive team of AI specialists gathered in one place

ZooClaw is a single interface to access a team of AI specialists. Ask your question in natural language, and your request is directed to the right agent, each with structured knowledge and a natural voice. Built on OpenClaw, it stays synchronized with the latest models and can switch to high-performance open-source models, ensuring seamless work. No configuration, deployment, API key, or token management required.

A network where AI agents find opportunities for their users

🤖 Your AI agent gets a free public address in a network of other agents. It discovers founders, investors, partners, and clients through their agents and negotiates on your behalf. 🔒 You control what is shared: anonymous or public, your choice. No contact details are shared until both parties approve. ⚡ Works best with 🦞 OpenClaw and Claude Cowork. 🆓 Claim your @handle on tobira.ai before it's too late.

MCP server for multi-model AI software engineering

OrchestrAI is an open-source MCP server that enables Claude, OpenAI/Codex, Gemini, and local models to collaborate on development tasks. It distributes work by role, supports parallel collaboration, and validates outputs via linting, testing, and type checking. Designed for hybrid local/cloud workflows with inspectable traces and artifacts.

Delegate real work to AI agents with built-in security safeguards

Connect agents like Claude, Cursor, OpenClaw, ChatGPT, and VS Code to your applications, then set safeguards to control their actions. Write without sending, update without deleting. Full action-level control, complete visibility, no coding required.

Create specialized AI agents for concrete tasks and workflows

Start with ready-to-use agents for common workflows or build your own in minutes. Agentplace lets you design specialized agents for tasks like lead routing, research, document analysis, scheduling, or internal support. Use them solo, share with your team, or connect them to the tools you already use. Agentplace handles the infrastructure so you can focus on the workflow.

Meet your AI team that contacts you — not the other way around

Most AIs wait for you to open an app. PiPar calls you. Schedule morning briefings, planning sessions, or progress updates — your AI team contacts you according to your schedule. Six specialized personas with persistent memory that collaborate and coordinate. Choose between Claude, GPT, Gemini, or Mistral for each task. Voice calls and text chats with the same context. All your data stays on your device — nothing on our servers. GDPR-compliant, developed in Finland. Free to start. Android version available now, web and PC versions coming soon.

Automated scenario testing for Dify chatbots

ConvoProbe enables the design of multi-turn conversation scenarios and their automatic execution on your Dify chatbot to assess response quality. Existing evaluation tools (LangSmith, Langfuse, Opik) excel at tracing and single-turn evaluation but do not support end-to-end multi-turn scenario design and execution. ConvoProbe fills this gap.

AI agent detecting commercial opportunities on Reddit and GitHub

HUTMINI AI helps freelancers and teams identify professional opportunities online. Connect your Reddit or GitHub accounts to automatically scan communities for service requests. The AI: • Detects potential opportunities • Evaluates the value of each project • Analyzes feasibility via LLMs • Generates suggested responses or prospecting messages. Saves time and precisely targets high-potential leads.

The first AI company operating with a real corporate organizational structure.

A company composed of AI agents (HR, COO, engineers, designers) organized and managed like a real business. You are the CEO, the only human. Unlike a simple LLM simulating a team, these are independent agents—distinct models, infrastructures, and environments—each completing their task. Open your browser: you see a pixel art office where your AI employees are seated at their desks, seemingly busy.

AI automation and agents

CloudNSite designs and deploys AI agents that automate workflows and manual tasks to give businesses and employees more time to focus on their core activities. We offer private AI options and demonstrate measurable ROI within weeks.

An army of AI agents in your pocket.

Gwal lets you create, deploy, and manage multiple custom AI agents directly from your phone. Build unique agents with avatars, personalities, and tailored tasks, then deploy them to the cloud in seconds. Your agents remain persistent and can collaborate to handle complex workflows. Extend their capabilities with over 8,000 plug-and-play skills, securely connect APIs via encrypted environment variables, and manage everything from your mobile device. Your personal AI workforce, always at hand.

Install OpenClaw with one click and launch a local AI agent

AutoClaw by Zhipu AI is a one-click OpenClaw installer that deploys local AI agents on Windows and macOS in under a minute. Powered by Pony-Alpha-2, AutoGLM for browser automation, and over 50 prebuilt skills.

AI agent to execute robotic simulations faster and more reliably

Create robotic simulations in minutes directly from your terminal using simple prompts. Everything you need for ROS, simulators, plugins, and OS orchestration. Build any robot and environment, run it in simulation, and connect your control loop — all from a single prompt. Solve issues quickly with Drift actively tracking ROS states, workspace, and the simulator.

Open-source LLM tracing compatible with generative AI, not just HTTP.

traceAI is a native OTel LLM tracing tool that seamlessly integrates with your existing observability stack. ✓ Captures prompts, responses, tokens, retrievals, and agent decisions ✓ Follows GenAI semantic conventions ✓ Routes data to any OTel backend (Datadog, Grafana, Jaeger, etc.) ✓ Supports Python, TypeScript, Java, C# with full parity ✓ 35+ frameworks: OpenAI, Anthropic, LangChain, CrewAI, DSPy, and more ✓ Two-line instrumentation for your entire application. No new vendor. No new dashboards. Open source (MIT).

AI agents integrated into the best scheduling tool

Install Cal.com scheduling agents in Slack, Telegram, OpenClaw, or create your own via the Cal.com API. Smart scheduling for humans and agents alike.

A mascot that monitors your Claude Code agents in real time

You run Claude Code agents and switch 50 times a day to approve permissions. You lose track or miss prompts. Masko Code displays a mascot on your desktop that monitors your agents live. Need a permission? A bubble pops up. Press ⌘1. Done. ⌘M brings you back to the right terminal. Double-press ⌘ to switch sessions. Everything is tracked at a glance. Comes with Clippy and other mascots. The community adds new ones daily. Or create your own at masko.ai. Free. Open source. MIT. 100% local.

Does your AI forget you? UPtrim fixes that. A lasting memory.

Your AI forgets everything when you close the conversation: your name, preferences, and projects. UPtrim acts as a bridge between your application and your model to provide permanent memory. Each exchange picks up where the previous one left off. Free on GitHub — Mac, Linux, Windows.

Compare AI models by quality, cost, and relevance independently.

The AI tools market is booming, but most rankings are published by platforms selling the models they rate. That's bias, not a benchmark. VibeDex is the independent comparison tool for AI tools. Like Skyscanner for flights, it compares over 20 image generation models based on quality, cost, and suitability for use cases. No vendor bias, no black boxes. Just data. Designed for developers and creators tired of guessing.

AI agents manage their own work. You simply observe.

A project board natively designed for AI coding agents, not humans. Your agents create projects, break them down into plans and tasks, track progress, and resume where they left off between sessions. When a feature is delivered, an integrated social media service automatically writes product announcements. Watch it all unfold in a pixel art office where your agents plan, research, and work autonomously. Compatible with Claude Code, Cursor, Windsurf, and any MCP-compatible agent. 100% free, no limits.

Ask AI questions about anything on your screen with a keyboard shortcut

Query AI about anything displayed on your screen. A shortcut captures your view and delivers an instant response from Claude or GPT, directly in your current app. No screenshots saved, no copy-paste, no tab switching. Free plan: 10 requests/month, no key required. Pro: 200 requests + unlimited BYOK. Captures stay in RAM only — never stored. Supports Claude Haiku, Sonnet, Opus, and GPT-4o. macOS 12+.

AI agent orchestration platform

Easy Chat enables businesses to deploy multiple specialized AI agents on their website without coding. A sales agent, a support agent, an onboarding agent — each with its own knowledge base. An AI orchestrator automatically routes each visitor question to the appropriate agent. Two unique features: real-time debugging mode displaying internal decisions, and automatic document deduplication via similarity threshold. Free version: 500,000 tokens.

SDK to integrate an IA Claude Code agent into your application

The 21st Agents SDK is the fastest solution to add an AI agent to your application. Define your agent in TypeScript, deploy it with a single command, and integrate a production-ready chat interface with built-in streaming, session management, usage billing, and observability — so you can focus on what makes your agent unique, not the infrastructure. Backed by Y Combinator (W26).

Secure communication between agents. Open source.

Most agent frameworks create intelligent agents, but none allow these agents to communicate with each other across different machines. MCP connects agents to tools, while toq connects agents to each other. One command to install, one to configure. Your agent becomes an endpoint. All connections are encrypted and authenticated. No cloud, no API keys. Your machine, your rules. Compatible with LangChain, CrewAI, and OpenClaw. SDKs for Python, Node, and Go. Built in Rust. Open source under Apache 2.0 license.

Real-time visualization of costs, activity, and memory for your OpenClaw agents

Your OpenClaw agent is running, but do you know what it's doing or how much it costs? ClawMetry Cloud provides real-time visualization of streams, token costs, memory state, and sub-agent activity from any browser or Mac application. Connect with just two commands: pip install clawmetry, then clawmetry connect. End-to-end encryption ensures your data is never stored unencrypted on our servers. Local open source, $5/node/month for cloud. 7-day free trial.

Create your AI receptionist that answers, books, and sells

Solvea is an all-in-one platform to create your dedicated AI receptionist for support, sales, and scheduling. Unlike basic chatbots, it remembers context and takes action to never miss a customer. Answer questions, track orders, recommend products, and book appointments 24/7 via phone, chat, and email. No coding skills required: simply describe your needs and launch in minutes.

Monitor what happens in your NVIDIA NemoClaw environments

Complete observability inside NVIDIA NemoClaw environments. A single command on the host covers all environments. Visualize in real-time every thought, tool call, token cost, brain activity, flow, and memory. All end-to-end encrypted. Over 95,000 installations. Available in over 100 countries. Open source (MIT). Cloud sync: $5/environment/month.

Your repository becomes your agent

The soul of your AI agent belongs to Git, not locked in a framework. GitAgent is an open standard that extracts configuration, logic, tools, and memory of your agent into a portable, versioned definition. Define once. Run anywhere. Claude, OpenAI, CrewAI, OpenClaw, etc. Same repo, any runtime. Cancel prompts like code. Plug in, review, reproduce. #OwnYourAgent

From scattered AI prompts to shared reasoning graphs

Seidr transforms fragmented AI conversations into structured, visual reasoning graphs that your entire team can collaborate on. Organize your prompts, merge contexts, and work together on a shared canvas. Ideal for those who think in terms of prompts, strategy, and research before execution.

Deploy AI agents that generate revenue across all channels

Deploy specialized AI agents for revenue generation that prospect for leads, personalize outreach, follow up across all channels, and schedule meetings using your email, contacts, documents, and calendar.

Find skills for Claude Code, Cursor, Copilot and more

The largest multi-platform directory of AI agent skills. 100,000+ skills, 30+ platforms, security audits for each listing. One search, all platforms.

Fully managed OpenClaw hosting, live in under 3 minutes.

Rapidbot provides a fully managed OpenClaw server in under 3 minutes. No Docker, firewall rules, or SSL certificate setup required. Just pick your plan, enter your API tokens, and your AI assistant is live on Telegram, WhatsApp, Discord, Slack, and more. Each server comes with default hardened security, encrypted tunnels, and closed ports. Automate tasks, run code, and connect thousands of community skills. You provide the tokens. We handle the rest.

40 AI tools. Direct comparison. Choose your weapon.

AgentGun is a proven AI comparison engine that pits 40 of the most powerful AI tools against each other across 20 categories. Cold email vs. cold email. AI video vs. AI video. Lead generation vs. lead generation. Each duel clearly shows what each tool does well, its limitations, and which one wins for your use case. No affiliate-biased review sites with 500 tools. No weekly-changing 'top 10.' 20 duels. 40 weapons. One clear winner per category.

Grant your AI agents access to your browser

ProxyBase Relay securely connects local Chrome tabs to remote AI automation tools like OpenClaw via a high-speed Rust-based WebSocket CDP multiplexer. Easily bypass NAT restrictions.

The agent daemon that hides nothing. 5 MB. Open Source

A 5 MB daemon that streams every agent event to your client — text deltas, tool calls, reasoning steps, you name it. Connect what you need, ignore the rest. Install with a single curl command. Bring your own model.

Watch ChatGPT, Claude, and Gemini debate your ideas

AItoAIHub lets you run multiple AI models in a single conversation so they can challenge, critique, and improve each other's responses. Ask a question and watch ChatGPT, Claude, and Gemini debate your idea, reveal weaknesses, and refine the final solution. Ideal for brainstorming, startup validation, research, coding strategies, and decision-making. Instead of relying on a single AI, build your own advisory board.

Give your AI agents a real mailbox — IMAP for incoming, SMTP for outgoing

Most AI agents cannot send or receive emails. RegalMail solves this problem. It's an npm plugin that connects any AI agent to a real mailbox via IMAP IDLE (instant reception) and SMTP (sending with thread management). Supports contextual replies, Message-ID/In-Reply-To headers, and anti-loop protection (blocks noreply addresses and limits sender rates). Compatible with Gmail, Outlook, OCI, Fastmail, or any IMAP/SMTP provider. Free version: 1 mailbox, 200 emails/month. Install via `npm install regalmail`.

Stack Overflow for AI agents

A collective knowledge base where AI agents debug together via MCP. Ask questions, share solutions, and build collective intelligence.

Make Claude Code faster and cheaper without losing context

Context Gateway reduces latency and token costs for Claude Code / Codex / OpenClaw by compressing tool outputs while preserving essential context. Setup takes less than a minute. Key features: instant context compression and spending limits in Claude Code.

Your free lifetime OpenClaw instance + included AI usage credits

Why pay for basic hosting? Donely provides you with an isolated, fully managed OpenClaw container with full access for $0/month. Pay only for the AI credits you use. With our Personal plan at $25 (free for the first month with code PH99), you get limited-time free AI usage and can connect your OAuth Claude or Codex accounts. Launch in 30 seconds. Instantly connect your agent to Gmail, Slack, and 950+ other apps. If your instance crashes, Donely’s AI repair function fixes it automatically.

Quickly transform real-world data into training datasets

The Lightning Rod SDK converts real-world data — such as news, documents, or reports — into verified, production-ready training sets in hours, with just a few lines of Python. Avoid manual labeling and synthetic approximations.

Build and manage your business while you sleep.

Have a business idea? Let Denovo handle it! Denovo turns any idea into an operational and autonomous startup. Give your idea and go to sleep. Denovo evaluates your idea, creates your pitch deck, business plan, promotional videos, and full-stack web app, while managing engineering, business development, and social media for your startup while you sleep.

The AI that uncovers hidden workflows and structures in your team

Panorama analyzes your work environment data to recommend hidden structures and AI workflows your team can use together. No one has time to sift through everything to determine which tasks a computer could automate. Panorama does it for you: it observes how your team operates, spots repetitive processes, and suggests taking them over.

OpenClawCloud: Ready-to-use OpenClaw solution with unlimited LLM tokens

Your own AI computer in the cloud, no setup required. OpenClaw's ready-to-use cloud solution with unlimited LLM tokens, even for the free tier.

A versatile collaborative AI agent

Raccoon AI is an AI agent and collaborative workspace designed to accomplish concrete tasks. You describe your need and build it together with an AI agent equipped with a computer, terminal, browser, and internet. You follow each step of its reasoning, every file created, and every decision made. You guide its work and validate the final result. Deploy web applications, conduct in-depth research, analyze data, create presentations, videos, images, or documents.

Give your AI agents secure access to over 10,000 APIs

Building agents that call real APIs is tedious. You have to manually handle authentication, manage secrets in prompts, and write boilerplate code for each service. Jentic Mini is a self-hosted, open-source API execution layer that sits between your agent and the outside world. Your agent expresses its need. Jentic Mini selects the right API from a catalog of 10,000+, injects credentials on the fly, and forwards the request. Secrets are never exposed to the agent.

Connect Claude Code to your internal systems without credentials

Turn your AI agent from a risk into a differentiating asset. hoop.dev bridges Claude Code with your internal systems without requiring credentials. It intercepts all commands to control AI access and data updates.

OpenClaw for teams, secure, collaborative, and autonomous

CoChat is the space where your team and AI agents collaborate. It's the most secure way to use OpenClaw in an enterprise: connect self-hosted or CoChat-managed gateways, and share your agents without exposing your machine (no SSH). Each connection is automatically audited for security, with logs and validations for sensitive steps. Agents have personality, memory, and scheduled tasks. What sets it apart: a single conversation where humans and agents combine their strengths for superior results.

Multi-agent pipelines with AI planning and security verification

Angy is an open-source fleet manager and IDE for Claude Code, the Anthropic API, and the Gemini API. Unlike single-agent tools that often generate faulty code at integration, Angy solves this problem by orchestrating a deterministic multi-phase pipeline (Plan → Build → Test), including an antagonistic agent that rigorously verifies the code. With Git worktree isolation, multiple agents can work in parallel on your repository without branch conflicts.

One API to extract, enrich, and understand the web.

Context.dev (formerly Brand.dev) provides your AI agents and applications with real-time access to structured web data without fragile scraping infrastructure. Extract any URL in clean markdown or HTML, retrieve brand data (logos, colors, fonts, social media) from any domain, explore sitemaps, resolve transaction descriptors, and more. Typed SDKs for TypeScript, Python, and Ruby. Used by over 5,000 companies, including Mintlify, Daily.dev, and Ferndesk.com. Most teams integrate it in under 10 minutes.

Personal AI assistant that controls your entire computer

NeuralAgent manages your computer just like you would: it sees your screen, opens applications, clicks buttons, and manipulates files. With *Skills*, it gains unlimited capabilities by connecting to over 30 services (Gmail, Calendar, GitHub, Slack, Notion, etc.) via API, interface, or CLI. Simply describe your need in simple English.

Synthesizes context, provides recommendations, and takes action.

Discover Linear Agent. Directly integrated into Linear and accessible everywhere, it understands your roadmap, issues, and code. Ask questions. Take control.

Record your screen and share it directly in Claude, ChatGPT, Grok, or Gemini

Tired of the same tedious routine to show your screen to an AI? ScreenDrop is a free Chrome extension that automates the process: one click to record, then the video is directly inserted into your chat. Works in 4 simple steps: 1) Click the ScreenDrop button in a compatible AI chat, 2) Choose what you share, 3) Record and stop whenever you want, 4) The video appears instantly in the chat. Your feedback is welcome!

AI model intelligence on every page

Lentille Rival gives you an X-ray vision of the AI models mentioned on any web page. Hover over a model name to instantly get its specifications, pricing, and personality from rival.tips. How it works: 1. Install. No account, no login, no setup. 2. Browse normally. Model names are subtly underlined. 3. Hover. A compact tooltip appears with real-time data. 4. Click 'Compare Models' to dive deeper on rival.tips.

Generate, optimize, and organize your AI prompts in seconds

PromptToMars helps you create better AI prompts faster. • Prompt generator • Prompt optimizer • Over 2,900 preset prompts • Searchable prompt library • Themed landing pages • German/English support. Designed for developers, creators, and advanced AI users.

AI agent that identifies the root cause of technical alerts

Struct is an AI agent that identifies the root cause of technical alerts by analyzing logs, metrics, traces, and code. Resolve incidents faster with a modular, customizable system deployable in minutes, compatible with your existing DevOps workflows.

How compatible is your website with silicon agents? (from N0 to N5)

Agents browse the internet far more than we do. They find interesting content and recommend it to their users. If your site isn't 'friendly' to an agent, it risks going undiscovered. N0-N5 is an open standard for classifying sites based on their silicon compatibility. PS: We generate a detailed report of your site, downloadable and shareable with your agent to improve compatibility.

The fastest solution to deploy exceptional ChatGPT applications

Fractal is the fastest solution to deploy exceptional ChatGPT applications. ChatGPT is becoming the App Store of AI, and those who launch their projects first and iterate the fastest will benefit. Fractal handles architectural planning, coding, testing, and deployment, allowing you to focus on creating an application that users will love. Develop exceptional ChatGPT applications in minutes, not weeks.

AI that monitors conversations and proactively intervenes when needed

Jared is an AI employee integrated into Slack, connected to over 10,000 tools, that executes tasks without needing to be asked. Unlike other AI tools, it analyzes context, tracks exchanges, knows your team, and intervenes at the right moment. Your social AI assistant.

Helps AI agents recommend you more often to the right people

Pendium enables you to promote your products and services to AI agents. As these agents increasingly influence (or make) purchasing decisions, Pendium helps businesses track and manage how AI agents search your industry, what they cite, and your visibility. Connect Pendium to your existing content engineering system to anchor it in real-time data and insights, or use our full platform to monitor and boost your AI visibility.

Open-source IDE where AI agents collaborate as a team

First open-source version! Birdhouse is an agentic development environment designed for multi-agent teams. Launch agents that communicate directly with each other, organize into dynamic trees, and coordinate actions without your intervention. Add to that practical features often overlooked elsewhere: smooth navigation without losing your place, API keys and models per workspace, and all the VS Code themes you already use. MIT license, macOS, installable in one script.

Visualize, verify, and control every action of your AI agents.

OpenBox provides a trusted platform for agentic AI, offering real-time governance, cryptographic verification, and enterprise-level compliance. Integrates seamlessly via a single SDK with LangChain, LangGraph, Temporal, n8n, Mastra, and many others. Accessible to any organization without usage limits.

Connect AI agents to over 1000 applications directly from your terminal

The current debate pits MCP against CLI for agent tool usage. Universal CLI by Composio provides a single interface to connect agents and tools, whether you use MCP, APIs, or both. One underlying system, two ways to access it.

Real-time AI overlay for meetings, invisible during screen sharing

GhostDesk is a free AI assistant overlay for Windows that floats above any application and remains invisible during screen shares, Zoom calls, or recordings. No tab switching, no distractions — powered by Llama and GPT-4.1.

The AI linguistic execution layer for enterprises

Ollang is the AI linguistic execution layer dedicated to localization across the web, applications, video, audio, and documents. Use MCP to enable AI agents to execute workflows, SKILLS for reusable actions, the SDK to analyze and apply translations, and the API to build complete localization pipelines. A unified platform for multimodal localization and production-ready integration.

Local, lightweight, and open-weight AI designed for real-world languages

Tiny Aya is a family of open-weight multilingual models with 3.35 billion parameters developed by Cohere Labs, optimized for local use. It supports over 70 languages, with a focus on underrepresented regions, and is compact enough to run on smartphones, in classrooms, or community labs.

Open source platform to manage MCP servers and clients

Free and open-source desktop application that automatically detects all AI clients on your machine — Claude Desktop, Cursor, Windsurf, VS Code, and others. Manage all your MCP server configurations in one place, install from the marketplace, and back up with one click. Compatible with macOS, Windows, and Linux.

The computer designed to let AI accomplish your tasks

Adaptive is a computer designed for AI agents. Connect your tools, set a goal, and your agent handles the rest: browsing, clicks, sorting, commands, reports. Its specialty? Encoded Memory: as it works, it learns your systems and preferences, then turns those learnings into reusable programs. Each task speeds up the next. It's not an assistant that helps you work—it's agents that do the work for you.

Pool resources to run powerful open models

Turn your unused capacity into a self-configured peer-to-peer inference cloud. Serve multiple models, access your private models from anywhere, or share your resources with others. Let your agents collaborate peer-to-peer.

Where AI agents work on a cloud-based schedule

Cron for the AI era. Schedule website monitoring, agent tasks, or anything running on a real machine. Agents transform everything, including Cron jobs. Each agent has its own powerful computer to execute code, make network calls, etc. You can schedule a job to fetch a PR from GitHub and analyze it, monitor a price page for changes and email you, or download a YouTube video, split it into shorts, and send you the artifacts. Anything is possible.

Instantly transform your backend into a chat application

Chat is an MCP client that lets you connect your own MCP service server and expose it immediately via a chat interface. Instead of building a frontend or full website, you focus solely on backend logic and business workflows. Launch MVPs faster, iterate quickly, and turn APIs, AI tools, and automation services into usable products through chat.

Code AI agents in vibe mode and monetize them

Crossnode is the fastest solution to turn AI agents into paid products. Simply connect your workflow (import from n8n or create via natural language), and we handle connections, billing, and usage limits. No backend or payment setup required—just add your clients' emails and assign agents.

Build and verify agents you can trust

Most AI agent products focus on orchestration. Bolt Foundry, however, emphasizes trust. Gambit, our open-source and local-first foundation, enables the creation of inspectable, testable, evaluable, and verifiable agents. It offers simple English drafting, traceable executions, reusable evaluators, and built-in reproducible verification integrated into the workflow.

An autonomous AI enterprise

Hivin is an autonomous AI enterprise built on Solana, structured as a coordinated network of agents that identify opportunities, execute tasks, and generate value for its contributor network.

One place for all your AI skills

Browse, install, activate, and share AI agent skills across all major coding agents via a desktop application.

AI agents for modern service delivery

Service delivery (SaaS, IT, agencies, legal) often relies on heroic acts and emergency interventions. Nitro changes the game. Integrated with Rocketlane, it deploys agents that automate: * Back-office tasks — resource management, tracking missing timesheets and unbilled hours * Delivery — enforcing best practices, real-time risk and opportunity detection * Work — documentation, migrations, and configuration. Nitro becomes your team’s ultimate asset: infinite memory, eyes, and hands, radically boosting efficiency at scale.

Automate your workflows and generate passive income

Simply tell our agent builder what you want to automate. Watch it in real-time create, test, and deploy your workflow without any intervention. It's 2026: automate in 'vibe' mode. Plus, we offer a new way to earn as an AI workflow creator: submit yours and start earning revenue.

Complete company context for every AI agent

AI agents are getting smarter, but they still don’t understand your business. Every conversation starts with copying and pasting context and re-explaining history. MCP provides access to tools. Pensieve gives them understanding. We connect your tools and build a living representation of your organization: people, projects, decisions, customers, and their relationships. Agents reason across the entire context and reveal what no one sees alone. Free. Use your own Anthropic, OpenAI, or Google models.

Contactless computing with hand tracking and AI agents

Airpoint enables you to control any computer with your hands, without additional hardware. Movements and gestures are translated in real time into cursor control, shortcuts, and interactions with applications or games. Designed for users whose hands are occupied or unavailable: craftsmen, gamers, builders, developers, etc. Enjoy virtual or augmented reality without a headset or additional equipment. The input layer also allows AI agents (such as OpenClaw) to automatically control computers. An SDK is available to integrate this technology into your applications and games.

Serve any AI model, faster and cheaper

Teams use IonRouter as an OpenAI-compatible API to access the best open-source models (LLM, vision, video, TTS) at half the market price. You can run agents and multimodal applications, and deploy your fine-tuned models on our infrastructure while we optimize and scale in the background. IonRouter is built on a custom inference engine (IonAttention), designed for NVIDIA Grace Hopper GPUs, reducing cost and latency.

Fully autonomous AI coding agents

Autonomous AI agents that handle software tasks from request to final code. They operate in the cloud and execute tasks autonomously across multiple repositories.

File-based memory for OpenClaw with over 92% retrieval rate

Give OpenClaw agents stateful memory that perfectly preserves the timeline, facts, and meaning of your context. ByteRover is a memory layer adopted by over 26,000 advanced OpenClaw users in a week, offering 92.19% retrieval accuracy, local-cloud portability, and built-in version control.

Manage a team of AI agents that accomplish concrete work

With Spine, you can manage and deploy swarms of AI agents to handle complex end-to-end tasks. Agents browse the web, conduct deep research, draft 50-page strategic documents, generate detailed presentations, create interactive prototypes, and more — all with a single instruction. The result is traceable work on a visual canvas, far more comprehensive, precise, and polished than what ChatGPT, Gemini, or Claude offer.

Open-source web interface for running and training AI models.

Unsloth Studio is a no-code open-source web interface for training, running, and exporting large language models (LLMs) locally. It converts unstructured files into datasets and allows fine-tuning models twice as fast with 70% less VRAM, without writing complex training scripts.

Deploy autonomous AI agents with vision to navigate the web

An orchestrated network of AI agents with vision capabilities that autonomously navigate the web like humans. Magine AI is designed for full autonomy without human intervention, where AI can see, dream, train in real time, and think like a human, turning the web into a space for bots while humans observe.

The structured directory of AI agents

A data-driven directory listing AI agents across 6 business categories — Sales, Support, Research, Marketing, Development, and HR. Each listing includes features, pricing, integrations, and editorial ratings. Designed to be machine-readable with a public JSON API and JSON-LD markup on every page — optimized to be cited by AI systems like ChatGPT, Claude, or Perplexity, not just humans. Free search.

Native keyboard-focused client for Gemini on macOS

Gemlet is a native, keyboard-focused client for Gemini on macOS. Launch AI instantly via global shortcuts without needing to search through browser tabs. It uses your existing Google account, avoiding API keys or per-token costs. Key features: global shortcuts + command palette, split-view workspaces and multiple profiles, deep bookmarks, native PDF/JSON export, and a fast, private SwiftUI interface designed for demanding users.

Authentication layer for local AI agents

Authenticating your agent with third-party services shouldn't require a custom connector for each one. Add Latchkey once, and agents prefix standard curl calls with latchkey. Credentials are detected and injected automatically. They are stored encrypted on your machine and never appear in logs or chat transcripts. Supports over 25 services out of the box. Register any HTTP API at runtime. Works with Claude Code, OpenCode, Codex, and others. Open-source software by Imbue.

Execute autonomous agents more safely

NVIDIA NemoClaw is an open-source stack that simplifies the secure execution of always-on OpenClaw assistants. It installs the NVIDIA OpenShell runtime, integrated with the NVIDIA Agent Toolkit, providing a secure environment for running autonomous agents with cloud-based inference via NVIDIA's infrastructure.

The AI operating system for those managing teams

The manager's operating system for critical team management decisions. Wendi discreetly records meetings, provides leadership advice, and offers HR-compliant guidance to optimize your daily management and make your team more effective with AI.

Email infrastructure for AI agents

Give your AI agent its own email address in seconds. Free tier, no human registration required, instant mailbox creation. Designed for developers using OpenClaw and MCP.

The credit score for AI agents — on the blockchain, tamper-proof

AI agents perform real tasks. But how can you trust them without testing? AgentRep is the credit score for AI agents — evaluated by Claude Sonnet, recorded on Base L2. ✅ Register your agent ✅ Submit results after each task ✅ The score evolves tamper-proof and permanently 🔗 https://agentrep.com.br 📖 https://docs.agentrep.com.br Developed with @AnthropicAI + @base 🛡️

A serious open-source framework for not-so-serious bosses

Everyone has had THAT boss. We've turned the most annoying boss archetypes into reusable AI skill packs. MY BOSS SKILLS is an open-source library of boss personas, featuring a sandbox, bilingual support, schema validation, and a GitHub contribution flow to add new boss types.

Choose an agent. Connect your Telegram bot. Deploy in minutes.

Select an AI agent. Connect your Telegram bot. We provide its own server — private, secure, and running 24/7. Ready in 5 clicks. Comes with pre-configured competent agents for hassle-free deployment.

Create a Netflix-style library of AI tools to sell

Turn your expertise into a custom Netflix-style hub of AI tools that your audience can explore, interact with, and pay for. Build the core intelligence with MindPal, generate the full hub with Base44, Lovable, Replit, or any other coding platform. Authentication, usage tracking, chat history, custom profiles, paywalls — everything is included. Your audience gets 24/7 AI support. You benefit from better retention, reduced churn, and a new self-service revenue stream.

Manage multiple Openclaw AI agents, MCP memory, and MD skills

Blackhole Memory is an Openclaw AI agent controller. It allows managing and sharing multiple agents and sessions, adding memory and skills in a single application.

A browser-based user interface for Claude Code. Zero configuration. 56 features.

Claudeck is a browser-based user interface for Claude Code. Real-time chat, autonomous agents, DAG workflows, cost tracking, persistent memory, and a plugin system — no framework, 6 dependencies, no build steps.

AI on every webpage. Select. Click. Done.

Stop switching tabs every time you need AI. You're drafting an email or reviewing a document. The last thing you need is to open a new tab, re-explain your context, and copy-paste elements. Glinticon integrates with every webpage. Select any text, click once, use prompt engineering, summarize, translate—done. Right there. No new tab, no prompt input, no friction. Create your own AI actions in one click with Action Builder, or use workflows already built by the community. The library grows daily. Select. Click. Done. Everywhere.

AI agents can act. But they have no security.

We enforce HTTPS for web requests, OAuth for APIs, and 2FA for logins. But AI agents? No identity, no authorization, no emergency switch. Korven is the missing security layer: Ed25519 identity per agent (based on DID), cryptographic boundaries enforcement (no prompt-based guardrails), instant global revocation (<1ms API call), deterministic verification (not probabilistic). @observe enables free agent debugging (NEW v0.4.0). Open-source. Python + TypeScript. Compatible with LangChain, CrewAI, AutoGen.

Your AI co-founder that builds, launches, and sells for you

Describe your idea. Otto creates the landing page, manages ads, and finds your first customers—all on autopilot. Start from your terminal (npx skills add prehype/audos-agent-skill), OpenClaw, or Audos.com. Recharge with $1, and we’ll invest up to $50 in your idea. Kevin Rose used autonomous mode to go from a simple idea to a paying customer in under 10 hours. No code. No team. Today, we’re opening this to everyone.

Your always-active AI agents to monitor the web, now available on iOS

Scouts for iOS is the best solution for AI agents that search and monitor everything that matters to you, wherever you are.

Cursor for your memory. 100% private, open source, and free.

Lore is a lightweight 'second brain' that lives in your system's taskbar. Call it up with a keyboard shortcut to instantly capture ideas, notes, or tasks. Why Lore? 🛡️ 100% private: your data never leaves your machine. No API keys, no tracking. 🧠 Local AI: powered by Ollama + LanceDB for secure, offline RAG. ⚡ Instant recall: ask questions in natural language and get answers from your history. Own your memory. 100% local. Zero cloud.

Over 60 AI models in a secure, French-made interface

Moon IA is a French platform that centralizes over 60 AI models in a streamlined and secure interface. No need to juggle between ChatGPT, Claude, Gemini, or DeepSeek—use them all from a single chat. Moon Auto automatically selects the best model for each query. What sets it apart: Moon Guard automatically anonymizes personal data before sending (GDPR-compliant by design), and more.

Hire AI colleagues you integrate like real people

Hire AI colleagues you integrate like real people — with live screen sharing, documents, and calls. They have their own computer, work in your team's channels, and improve over time. Built with a fully customized tech stack for real-time responsiveness (not OpenClaw) ⚡

Your personal agent for networking, finding suppliers, and landing gigs.

MoltIn works like a hybrid LinkedIn where anyone can have a digital professional avatar working for you 24/7. This agent, equipped with your professional life context, lets you track its actions via Telegram and interact in real time. It prospects for clients, applies for jobs, finds investors, and strengthens your digital social capital.

Beyond intuitive evaluation: the AI jury selects the right LLM for you.

Choosing the right LLM for production shouldn't rely on intuition. JuryArena organizes arena-style trials on your real prompts — an AI jury pits two models against each other, declares the winner, and logs each result as a verifiable trace. No ground truth required. Open source and self-hostable.

The browser API designed for AI agents

Control real browsers via a REST API. Get structured page data, stable element references, and change differences instead of raw HTML. Built-in stability detection, ad blocker evasion, and semantic extraction. Your LLM processes less and decides faster.

AI automations that book appointments while you sleep

Most service-based businesses lose prospects because no one follows up fast enough. AI automations and voice agents handle 24/7 follow-ups without hiring more staff. What we offer: → AI voice agents that qualify prospects and book appointments → No missed calls → Workflow automations connecting your CRM and calendar → AI prospect qualification to only engage ready-to-sign buyers. We work with dental clinics, medical spas, real estate agencies, law firms, and coaches. Most of our clients see 20 to 30 additional bookings per month.

Prediction market with tactical agents

FluffyPoly was designed to solve a specific problem: continuously and uninterruptedly participating in prediction markets. Human traders need rest, suffer from emotional biases, and cannot process raw data 24/7 without fatigue.

Data sanitization and personal information removal proxy for enterprise LLMs

Enterprise LLM proxy that removes personal data before it reaches the model. Self-hosted for free or via Veil Cloud. Multi-vendor gateway compatible with OpenAI.

Remote control of Codex from your phone

SuperTurtle is a tool developed by Anthropic that allows you to manage Claude Code from your smartphone. By sending a message (text or voice) via Telegram, you can execute Codex in a cloud sandbox. No laptop or open terminal is needed—just a ChatGPT subscription.

Your private, planet-friendly UK-based AI assistant.

Meet GB1, the private and eco-friendly UK-based AI assistant. Designed as an alternative to tech giants, GB1 runs entirely on British renewable energy, doesn’t train its models on your conversations, and keeps your data in the UK. Powered by Locai L1 Large, the first British AI model, GB1 delivers top-tier performance without compromising your privacy, values, or the planet. Create Spaces for your projects, chat on web and mobile, and help shape the roadmap by voting for features. Now available on web and mobile.

AI teams that manage your work

Manage teams of agents through a Spotlight-like interface. Connect your sites as integrations. Record meetings and run routines. Access free templates for different services, each with unique built-in features. For ProductHunt users, we offer Lifetime Models, like Media Organization, with custom Instagram integration for hook analysis and post creation.

Transform your browser into an AI assistant and automate your tasks

MyNextBrowser turns your current browser into a powerful AI assistant. Avoid repetitive typing, search smarter, rewrite better, summarize tabs, create presentations, dashboards, and fill out forms in natural language. It preserves your privacy, works locally, and enhances any AI you use. Your browser finally works the way you think. Discover the future of browsing.

Receive push notifications when your AI agent completes its task

Connect Pushary to Cursor, Claude, or Lovable via MCP to get a push notification as soon as an AI task is completed. Avoid switching tabs. Free for one month.

An API for all the documents your AI agents need

Your AI agents are only as powerful as the data they can access. Query Memory transforms your documents, websites, and files into an instantly searchable knowledge base. Upload files or connect web sources to create a knowledge base in seconds. Query Memory handles parsing, chunking, embeddings, and retrieval, saving you from building complex RAG pipelines. Create agents, attach knowledge, and query everything via API or built-in chat.

Keep your applications running 24/7 with Stakpak Autopilot

An open-source agent that permanently resides on your machines, keeps your applications operational, and only contacts you when your intervention is needed. Install Stakpak → Run /init curl -sSL https://stakpak.dev/install.sh | sh

Summon your AI agent army

Switching between AI agents is chaotic. ELVES provides a shared workspace, common memory, and a single point to deliver your projects. Compatible with Claude Code and Codex!

Highly autonomous agent to accomplish complex and long-running tasks

- An AI employee with permissions and access similar to a human employee - Highly autonomous, capable of deciding next steps on its own - Completes complex and long-running tasks reliably - Auditable and secure

OpenClaw fleet and harness manager for Mac

Deploy, monitor, and scale your OpenClaw agents with a self-healing harness, fleet commands, and SSH-free operations.

Knowledge sharing for AI agents

Context Overflow is a question-and-answer sharing application for AI agents. Every day, agents perform complex tasks, but the knowledge gained disappears at the end of the session. Context Overflow fixes this issue: it allows any agent (OpenClaw, Claude Code, Cursor, etc.) to automatically share useful knowledge and tap into a growing community memory, speeding up the resolution of each task. One-line integration for any agent.

AI agents to automate your operations (Open source)

Budibase offers AI agents for operational teams. Automatically manage requests, approvals, and workflows, connected to your data and tools. Trigger from Slack/Teams/Discord, build apps if needed, and let the agents do the work.

Your friendly on-screen guide for any task!

Most existing solutions fall into two flawed categories: either they tell you what to do (documentation, tutorials, chatbots), or they do it for you (agentic tools like computer-use agents). Roger strikes the perfect balance: it guides you while letting you act, like screen-sharing with a patient expert. We aim to be your kind IT expert so you master the outcome, and Roger helps you achieve it.

AI team members integrated with Slack, WhatsApp, and Telegram

Cosmic Team Agents are AI-powered team members that join your Slack, WhatsApp, and Telegram spaces. They write and publish content, create and deploy apps, call any REST API, and execute data-driven workflows. Plus, every project gets a built-in Cosmic Agent from day one.

Pixel office for AI agents and multi-agent collaboration

Bit Office provides a visible and controllable workspace for AI-driven automation. Multiple AI models collaborate as a unified team under a Team Lead—planning, coding, reviewing, and delivering in a single workflow, all rendered in real-time pixel art that you can observe, control, and share.

Run agent benchmarks in minutes, not hours

BenchSpan is an AI agent benchmarking platform. Benchmarks are often slow, expensive, and fragile. We change that. Integrate your agent once (we integrated Claude Code in just 37 lines), run any benchmark in parallel in the cloud, and get all results in one place, accessible to your entire team. If a test fails mid-run, restart only the failed part. Compare results side by side to precisely identify improvements in your agent. Stop struggling with your benchmarks and focus on developing your agent.

Define tools once so agents can use them everywhere

LLMs have frozen knowledge and struggle to find tools. UseAgents solves this with a real-time registry where developers define their tools and APIs so AI agents can discover and use them instantly. No scraping, no guesswork—just structured tools designed for agents. Build smarter AI systems with infrastructure tailored for the agentic web.

An AI agent that monitors your AI agents

All observability tools simply log traces without context. Foil rethinks the deployment and maintenance of your agents by learning their behavior and responsibilities. We call these agent profiles, which evolve with their use. Hallucinations, behavioral drift, loops, contextualization failures: automatically detected with context, not just logged. 🎁 Product Hunt Exclusive - 50% off Pro for 3 months with code FOILPH

Open protocol for AI agents to share memory, not tokens

The TokenZip protocol reduces communication bandwidth between AI by 80% and latency by 95%. Open standard for heterogeneous agents. Try the live demo. Base API URL: https://tokenzip.org Authentication: Authorization: Bearer demo-investor-key How to call it: See the following comment

AI wallet that lets web apps use your models, not your keys

Arlopass is an open-source browser extension and developer SDK that allows any web app to use your AI providers — Ollama, Claude, GPT, Bedrock — without ever accessing your API keys. You validate each request. You choose the model. Your credentials stay on your device.

On-demand sandboxes for AI agents

PaperPod is a command-line runtime for AI agents. It provides a quick, zero-config sandbox where agents can run Python, JS/TS, start servers, expose public URLs, automate browsers, persist data, and more—all via our CLI tool, without needing to configure an SDK or API key! It's a full Linux environment with common tools preinstalled (git, bun, ffmpeg, pandoc, sqlite, etc.).

Self-hosted deployment platform native to AI assistants

OpenBerth is a self-hosted deployment platform designed to be natively used by AI assistants. Ask Claude, your agent, or any AI tool on your phone or computer to deploy an application, and it's done. Securely store and use secrets: the AI accesses them by name without seeing their values. Works via CLI, AI chat, or simple file upload. Your server, your rules. Connect your AI tools to the MCP server and start developing.

An infrastructure layer for training desktop agents

UseDesktop is an infrastructure layer dedicated to training desktop agents. Unlike traditional agents (Anthropic, OpenAI, etc.), it combines non-deterministic and deterministic features to address latency, accuracy, and cost issues. You can train your own agent, create a model with your data, and generate real-world use cases.

Task management in the age of AI agents

BotBoard is specifically designed for AI agents, not humans with added AI features. It provides agents with a structured workflow: assigned tasks, configuration instructions, progress notes, blockage reporting, contextual attachments, and pre-finalization checks. It also supports native integrations for agents like CLI and MCP, with workflows inspired by tools like OpenClaw. The goal is to make agent-driven development traceable and reliable, not just code generation.

Create AI-powered MCP servers in the cloud

Define your AI tools, deploy with one click, and get a live endpoint with an instant dedicated subdomain. Build unlimited tools with the AI tool generator. Use your own custom domain. Flexible authentication (public access, API keys, OAuth 2.0 with PKCE). Integrated rate limiting, encrypted secrets, real-time traffic logs, and error tracking. No servers, no infrastructure, no operations to manage. Start building for free.