All AI agents

OpenAI's coding agent, offered as an open-source terminal CLI, an IDE extension, a desktop app and a cloud agent that works on repositories. Users sign in with a ChatGPT account or use an OpenAI API key; with an API key, Codex works in the CLI, SDK and IDE extension without cloud features such as GitHub code review and Slack. On the Free and Go plans Codex runs in the desktop app (subject to rollout); the web, CLI and IDE extension come with Plus and higher.

Anthropic's agentic coding tool that reads a codebase, edits files and runs commands from the terminal, with VS Code and JetBrains integrations plus desktop, web and mobile access. It runs Claude models through a Claude subscription, an Anthropic Console API key or supported third-party cloud providers.

A workflow automation platform with a visual editor, several hundred integrations and native AI agent nodes. It can be self-hosted or used as a cloud service.

A framework for building LLM applications and agents from composable parts: model integrations, tools, retrievers and prebuilt agent abstractions. Available for Python and JavaScript.

Open-source coding agent from Anomaly (the team behind SST) that runs as a terminal UI, desktop app or IDE extension and connects to 75+ model providers, including local models and existing ChatGPT or GitHub Copilot logins.

OpenShell is a runtime that runs AI agents in isolated sandboxes and enforces declared policies on file access, system calls and network connections, injecting credentials only for approved endpoints; it ships a CLI, gateway and SDKs.

Symphony is a specification and experimental Elixir reference implementation from OpenAI that watches an issue tracker such as Linear and spawns Codex agents to implement tasks, returning CI status, PR reviews and walkthroughs for human review.

Open-source personal AI assistant that runs on the user's own machine as a gateway and works through chat apps such as WhatsApp, Telegram, Slack, Discord and iMessage, with a skills system and swappable model plugins. It also ships native apps for macOS, iOS, Android and Windows. Its README warns that tools run on the host for the main session unless sandboxing is configured, and that inbound messages should be treated as untrusted input.

Minimal, extensible open-source coding agent harness with interactive TUI, print/JSON, RPC and SDK modes, created by Mario Zechner and now maintained by Earendil Inc. It supports 15+ model providers through API keys or OAuth and is customized with TypeScript extensions, skills and packages. Its tools and extensions run with the permissions of the Pi process; Pi does not sandbox tool calls or ask before each one.

Codex Security is OpenAI's application security agent; this CLI and TypeScript SDK (@openai/codex-security) scan repositories or Git changes, validate findings, generate patches and export results such as SARIF. Running scans requires Codex Security access on an OpenAI account. A separate Codex Security Cloud plugin (research preview) scans connected GitHub repositories in Codex cloud.

AI coding editor and agent platform with a desktop app, a CLI, web and mobile agents, and cloud agents that run tasks in remote environments. It offers models from OpenAI, Anthropic, Google, SpaceXAI (Grok) and Meta alongside its own Composer models.

An AI gateway and Python SDK that calls 100+ model APIs through one OpenAI-style interface, with spend tracking, budgets, rate limits, fallbacks, guardrails and logging.

A low-level orchestration framework for building stateful agents as graphs, with durable execution, persistence and human-in-the-loop steps. Available for Python and JavaScript.

AI coding assistant from GitHub offering completions, chat and agent mode in IDEs, a terminal CLI, a desktop app and a cloud agent that works on assigned tasks in the background and opens pull requests. Models from several providers can be selected.

NemoClaw is a reference stack and CLI that installs and runs agents such as OpenClaw, Hermes and LangChain Deep Agents Code inside NVIDIA OpenShell sandboxes, with guided onboarding, managed inference, network policies and state persistence. It is an alpha project.

The Unity CLI includes an MCP server, started with `unity mcp` and built on the Unity Pipeline package, that lets external coding agents such as Claude Code, Cursor, VS Code and Windsurf control a running Unity Editor locally. It replaces the deprecated MCP server in the com.unity.ai.assistant package, is marked experimental, and works alongside Unity's skills-only plugin for Claude Code, Codex and Grok.

Unity's agentic assistant runs inside the Unity 6 Editor, inspects the scene, GameObjects and components, carries out Editor actions and checks that the changes work; it also includes asset generators that an organization can switch off. Unity is retiring the "Unity AI" brand name, the tools are in beta, and they are meant for Unity developers.

Vercel's TypeScript toolkit for building AI applications and agents, with one API across model providers and UI helpers for streaming.

An open-source platform for tracing, evaluating and debugging LLM apps and agents: every model and tool call with its cost and latency, prompt management, datasets and evaluations.

Turns websites into clean Markdown or structured data for AI agents: scrape, crawl, map and search the web through one API.

Open-source self-improving agent from Nous Research that keeps persistent memory, creates and refines skills from experience, runs scheduled automations, and works from a terminal interface or messaging apps such as Telegram, Discord, Slack and WhatsApp. It can run locally or on Docker, SSH, Modal and Daytona backends, and has a desktop app.

Open-source software-engineering agents from Princeton and Stanford researchers: SWE-agent (2024) and its roughly 100-line successor mini-SWE-agent, which uses only bash and works with any model through LiteLLM and similar routers. The SWE-agent README says development effort has moved to mini-SWE-agent.

An open-source SDK from AWS for building and controlling agents in Python and TypeScript.

An open-source platform for building agentic workflows, retrieval pipelines and LLM applications with a visual builder. Available self-hosted or as a cloud service.

High-performance open-source code editor written in Rust, with multiplayer editing, an agent panel, edit predictions and support for external agents through the Agent Client Protocol. Version 1.0 shipped in April 2026.

Anthropic's SDK that exposes the agent loop, tools and context management behind Claude Code, for building custom agents in Python and TypeScript.

Stockholm-based prompt-to-app builder that generates full-stack web apps with a built-in cloud backend and Git sync, and can be used from iOS/Android apps, Slack and Telegram. It grew out of the GPT Engineer App.

Secure cloud sandboxes where AI agents run code, use a filesystem and terminal, and browse, isolated from your own machines.

An agent harness from DeepSeek in which every capability is a plugin. Its agent reads and edits workspace files, runs commands, delegates work and keeps a plan, and it runs as a desktop app, a browser UI or the dsh command-line tool.

OpenSRE is an open-source framework and CLI for building AI SRE agents that connect to existing infrastructure tools to investigate incidents, diagnose failed builds and open fixes as pull requests. It can run self-hosted with the user's own model provider; per the README, the first launch of the CLI activates the hosted model and the interactive shell requires an OpenSRE account. A managed OpenSRE Cloud gateway, provisioned by an organization admin in the web app, is also available.

Framework and platform for building voice and chat assistants that can be self-hosted: the open-source Rasa framework (Apache-2.0) and the open-core Rasa Pro / Rasa Platform. The new Rasa Platform centres on the Mantle orchestrator, which builds agents from skills written in natural language; its agent-building UI is in early access (Rasa docs). A free Developer Edition license covers one bot with capped monthly conversations.

OpenAI's lightweight framework for multi-agent workflows, with handoffs, guardrails, sessions and tracing. Available for Python and TypeScript.

Google's open-source terminal agent that brings Gemini models to the command line for code editing, shell commands and MCP tools. Since 2026-06-18 it no longer serves free, Google AI Pro/Ultra or Google One users, who were moved to Antigravity CLI.

AI agent inside Replit's browser-based platform that builds, tests and deploys full-stack web and mobile apps from natural-language prompts. It runs in Free, Power and Max agent modes and includes hosting, databases and a design product.

Open-source Python library and hosted cloud that let LLM agents control a real browser to complete web tasks, with a CLI, a cloud API for hosted agents and browsers, and its own BU 2.0 model option.

Open-source Node.js server and React app for running a team of AI agents toward a business goal. You define the goal, give agents such as Claude Code, Codex, Cursor, Gemini CLI, OpenCode or OpenClaw roles in an org chart, set a budget per agent and approve the plan, then follow tasks, costs and an append-only audit log from one dashboard. A hosted Paperclip Cloud is on a waitlist.

A Python framework and runtime for building multi-agent systems with memory, knowledge and tools. Formerly known as Phidata.

A batteries-included agent harness built on LangGraph, with planning, a virtual file system and sub-agents for long-running tasks.

Open-source terminal coding agent (`omp`), built by Stencil Labs as a fork of the Pi coding agent by Mario Zechner, with a native Rust core, IDE-style tooling and Windows-native support.

A framework for programming language models with declarative modules and optimizers instead of hand-written prompts.

ElevenLabs' conversational agent platform (formerly Conversational AI) for building and deploying voice and chat agents across web, mobile, telephony and messaging channels.

JetBrains' LLM-agnostic coding agent, available inside JetBrains IDEs, as a terminal CLI, in JetBrains Air and in CI through GitHub and GitLab integrations. It supports bring-your-own-key models and a free Junie Lite tier.

Open-source autonomous coding agent for VS Code (and JetBrains) that also ships a CLI, a desktop app and an SDK. It works with many model providers through the user's own keys, Cline's usage-billed provider or the ClinePass subscription.

Browser Harness | Self-healing harness that enables LLMs to complete any task.

Cloudflare's SDK for building stateful agents that run on Workers and Durable Objects.

A TypeScript framework for building agents and AI workflows, with memory, evals, retrieval and observability built in.

Skill for Claude Code, Cursor, Codex and Gemini CLI: typing /graphify maps a project's code, docs, PDFs, images and video into a knowledge graph the agent queries instead of searching through files. Code is parsed locally with tree-sitter, without an LLM, embeddings or a vector store, and every edge is marked as extracted from the source or inferred. Graphify Cloud keeps the graph live and adds PR review and formal verification.

Connects AI agents to 1,000+ apps and APIs with managed authentication, tool search and a sandboxed workbench, so agents can take actions in tools like GitHub, Slack and Gmail.

Google's agent-first development platform, introduced as an IDE based on a VS Code fork alongside Gemini 3. Antigravity 2.0 (2026-05-19) added a standalone desktop app for orchestrating parallel agents, next to the IDE, a CLI and an SDK.

A full-stack SDK for agent-native application frontends with chat UI, generative UI, shared state and human-in-the-loop for React, Angular, Vue, React Native and chat channels; the team maintains the AG-UI protocol.

Native Rust command-line tool that lets AI agents control Chrome through compact accessibility-tree snapshots and ref-based commands such as open, click and fill. It is installed via npm, Homebrew or Cargo.

General-purpose autonomous agent that plans and runs multi-step tasks in a cloud sandbox, such as research, slides, websites and data analysis, with web, desktop, mobile and browser-extension clients. Manus 2.0, announced in September 2026, added a new agent harness, dedicated cloud computers and a desktop studio.

Open-source SDK from Browserbase for building browser agents that mix natural-language actions (act, extract, observe, agent) with code, available in TypeScript, Python and Go and runnable locally or on Browserbase cloud browsers.

Rust-based terminal that has grown into an agentic development environment, with a built-in coding agent, a standalone Agent CLI and the Oz cloud-agent platform. Its client was open-sourced in April 2026.

MCP for Blender is an open-source add-on and MCP server that lets LLM clients such as Claude, Codex, Cursor and VS Code control Blender, editing scenes and pulling assets from Poly Haven, Sketchfab and Poly Pizza. It can also generate 3D models through Hyper3D Rodin, Hunyuan3D or Tripo, and it is an independent project not affiliated with the Blender Foundation.

A Python framework for orchestrating teams of role-playing agents (crews) and event-driven workflows (flows).

Give your AI agent eyes to see the entire internet. Read & search Twitter, Reddit, YouTube, GitHub, Bilibili, XiaoHongShu — one CLI, zero API fees.

Chatwoot is an open-source omnichannel support desk (live chat, email, WhatsApp and social inboxes) and Captain is its AI agent that answers from help-center content, PDFs and websites, and also drafts and suggests replies for human agents. The core is MIT-licensed with an enterprise directory under a separate license, and it can be self-hosted or used as a hosted plan.

A framework for building realtime voice and video AI agents on LiveKit's WebRTC infrastructure.

Google's open-source, code-first toolkit for building, evaluating and deploying AI agents, with SDKs for Python, TypeScript, Go, Java and Kotlin.

Agent skill that creates a 20-second launch video with music, motion and share copy for the project in the current repository. It runs in Claude Code, Codex and other skill-compatible agents; a hosted service, letsbrag.app, makes the video from a website URL.

A memory layer for AI agents and apps: it extracts facts from conversations, stores them and retrieves the relevant ones later, so an agent remembers users and past work across sessions.

A Python agent framework from the Pydantic team, built around type-safe structured outputs, dependency injection and model-agnostic tool calling.

Orca is an open-source desktop agent development environment that runs Claude Code, Codex, OpenCode, Pi and other CLI agents side by side, each in its own git worktree, with terminals and a mobile companion app for monitoring agents.

Open-source add-on for Claude Code, Codex, OpenCode, Cursor and other coding agents: skills, commands, hooks, rules and memory that make the agent plan before it builds, verify changes with tests, review its own work from a fresh context and turn repeated wins into reusable skills. Installs as the ecc@ecc Claude Code plugin or through the npm package ecc-universal; ECC Pro is a hosted GitHub App for private repositories.

A lightweight, self-hosted personal agent runtime in Python that runs in a web UI, a terminal or chat apps, using shell, files, web, MCP and sub-agents, with long-term memory and scheduled automations.

AWS's AI coding assistant (the successor to Amazon CodeWhisperer) offered as IDE plugins, a command-line agent and chat in the AWS console, Slack and Teams. AWS is retiring the IDE plugins in favor of Kiro. Winding down

An open-source web crawler and scraper that turns any website into clean, LLM-ready Markdown, run on your own machine or through Crawl4AI Cloud.

A visual builder for AI agents and LLM workflows, distributed as an npm package and a Docker image.

A self-hosted personal assistant built on AgentScope that runs locally, in Docker or in the cloud, with skills, plugins, MCP, multi-agent support, layered memory and scheduled tasks, and connects to chat channels such as DingTalk, Lark, Discord and Telegram.

An all-in-one AI application for chatting with documents, multi-user workspaces and built-in agents, including a no-code agent-flow builder and MCP compatibility; runs as a desktop app or self-hosted with Docker.

A framework for building LLM applications and agents over your own data, covering document processing, retrieval and agent workflows.

Autonomous AI software engineer from Cognition that runs in cloud VMs, takes work from Slack, Teams, Linear, Jira or an API, and delivers pull requests. Devin Desktop and the Devin CLI act as local companions to the cloud agent.

Open-source system that turns an AI coding assistant into a video production studio: you describe the video in plain language and the agent researches, writes the script, generates or gathers assets (including free stock footage and open archives), edits a timeline and renders the result, showing a contact sheet with per-asset cost before the final render. The project lists 12 production pipelines, 100+ tools and 700+ skill and production-knowledge files.

A Python library for anomaly detection across tabular, time series, graph, text, image, and audio data. 60+ detectors, benchmark-backed ADEngine orchestration, and an agentic workflow for AI agents.

Local-first search engine and MCP server for AI agents. One Rust binary indexes code, docs, logs, JSONL/CSV and PDFs for BM25 full-text, vector and hybrid search, code search without grep, agent memory and RAG. Elasticsearch-compatible API; Corpus Hub serves hundreds of reference memory corpora.

Open-source, general-purpose AI agent that runs locally as a desktop app or CLI and can write code, run commands and use MCP extensions with 15+ LLM providers. Created by Block, it is now hosted by the Agentic AI Foundation at the Linux Foundation.

AWS's agentic IDE built around spec-driven development, with a CLI, Kiro Web for cloud sessions, a mobile app and the open-source Kiro Crew workspace. AWS positions it as the replacement for Amazon Q Developer's IDE plugins.

Understand Anything is a plugin for Claude Code and other coding agents that analyzes a codebase or document wiki with a multi-agent pipeline and builds an interactive knowledge graph of files, functions, classes and dependencies, with search, plain-English summaries and guided tours.

Compresses tool outputs, logs, files, RAG chunks and conversation history before they reach the model, so agents use fewer tokens. Runs locally as a library, a proxy, an agent wrapper or an MCP server; originals stay cached and can be retrieved on demand.

Platform for building, testing, deploying and monitoring AI voice and chat agents, with built-in telephony, prompt and flow builders, tool calling and analytics.

Platform for deploying AI agents that resolve customer support conversations over chat, email and voice, with workflows written as natural-language Agent Operating Procedures. It also provides testing, QA and analytics tooling for those agents.

A TypeScript SDK, runner and CLI for end-to-end tests of web and mobile apps in which agent steps carry out natural-language tasks in real browsers, simulators or emulators; agent steps are recorded on the first run and replayed without model calls until the app changes.

A smarter, self-hosted AI assistant — multi-user, multi-agent.

OpenViking is an open-source context database for AI agents that stores knowledge, memory and skills as a virtual filesystem under viking:// URIs, with scoped semantic search and generated summaries agents read before full content.

A filesystem-first framework for durable backend AI agents in which instructions, tools, skills, channels and schedules live as files in conventional locations; created by Vercel and in beta.

Tool that generates command-line interfaces for existing desktop software, codebases and web APIs so AI agents can operate them, plus CLI-Hub, a registry and cli-hub package manager for community-built CLIs. It works with Claude Code, Codex, OpenClaw and other skill-compatible agents.

AI workspace built around a Super Agent that plans and runs multi-step tasks in a sandbox with a real browser and filesystem, producing slides, sheets, documents, research and websites. It also offers a Chrome extension, an AI browser, a Claw personal agent that can run in the cloud or locally, and in-house models such as Gen-1.

Open-source memory system for AI agents from Vectorize that stores what an agent learns and recalls it in later sessions.

MCP for Unity is an open-source bridge that lets AI assistants such as Claude Code, Cursor, VS Code, Windsurf, Cline and Gemini CLI manage assets, control scenes, edit scripts and automate tasks in the Unity Editor. It is a Unity package (git URL or OpenUPM) plus a Python MCP server, and its README says it is sponsored and maintained by Aura.

Open-source AI software-development agent platform (formerly OpenDevin) with a local web GUI, a terminal UI/CLI, a hosted cloud and an SDK. It is model-agnostic and can use the user's own keys.

Open-source agent meta-harness for Claude Code and Codex by rUv: it installs as Claude Code plugins (core tools, a visual console and mods that add routing and policy enforcement) or runs as an MCP server for Codex and other MCP clients, and coordinates teams (swarms) of agents with shared memory, self-learning and vector search.

deepset's open-source framework for building production LLM pipelines and agents, with a focus on retrieval.

Alibaba's open-source terminal coding agent for Qwen models that also offers a desktop app, web UI, IDE plugins (VS Code, Zed, JetBrains) and SDKs. It accepts OpenAI-, Anthropic- and Gemini-style provider APIs as well as local models.

Developer platform and API for building, testing and deploying voice AI agents for phone and web calls, combining third-party speech-to-text, language model and text-to-speech providers.

GitHub's standalone terminal coding agent (`copilot`, npm @github/copilot) that plans and edits code, runs commands and works with GitHub from the terminal. It runs on a Copilot plan and also supports bring-your-own model providers.

Open-source coding agent available as a VS Code extension, JetBrains plugin and CLI, with a gateway to 500+ models and support for bring-your-own keys and local models. Kilo launched in March 2025 and was acquired by Anaconda in July 2026.

AI coding agent and chat for VS Code and JetBrains IDEs with an agentic planning mode, multi-file edits, MCP support and custom agents. Works with Claude, GPT, Gemini, Grok and other models, your own API keys, or local models via Ollama and LM Studio.

Traces, evaluates and troubleshoots LLM apps and agents, built on OpenTelemetry, with prompt experiments and datasets. Runs locally, self-hosted or in the cloud.

A visual, Python-based builder for AI agents and workflows that can also expose flows as APIs and MCP servers.

Enterprise AI agent platform for customer service that separates language handling from a deterministic Decision Engine, so refunds, claims and identity checks run by policy. It deploys one agent across voice, chat, app, email and contact-center channels and has an open agent SDK.

CodeGraph parses a codebase with Tree-sitter into a local, queryable knowledge graph that stays in sync with code changes and exposes it to AI coding agents such as Claude Code, Codex and Cursor over MCP.

Gives coding agents up-to-date, version-specific documentation and code examples for libraries, through an MCP server or API, instead of relying on what the model memorised.

Microsoft's open-source framework for multi-agent conversations and event-driven agent systems.

codebase-memory-mcp is an MCP server that indexes a codebase into a persistent local knowledge graph using tree-sitter and serves structural queries to AI coding agents. It ships as a single native binary for macOS, Linux and Windows.

Open-source desktop AI client for Windows, macOS and Linux with chat, autonomous agents and 300+ ready assistants, connecting to many cloud and local LLM providers.

An open-source Python framework for voice and multimodal conversational agents, maintained by Daily and the community.

AI code assistant and agent delivered as IDE plugins (VS Code, JetBrains, Visual Studio), a CLI and an OpenCode plugin, with a context engine and private or self-hosted deployment for enterprises.

Builds real-time knowledge graphs for AI agents from conversations and business data, tracking how facts change over time, so agents can query what is true now and what was true before.

Platform for building and deploying LLM-based support agents with a hosted studio, a conversation-metered pricing model and a public MIT-licensed monorepo of SDKs and integrations. It integrates with help desks such as Intercom, Zendesk, HubSpot and Salesforce.

code-review-graph parses a codebase with Tree-sitter into a local, incrementally updated graph and serves compact review context to AI coding tools over MCP and a CLI. An install command configures tools such as Claude Code, Codex and Cursor.

Packs a whole repository, or the parts you choose, into one AI-friendly file with a token count per file, so you can give a codebase to a model or agent in one go.

AI search engine for peer-reviewed scientific literature. Its Deep Search mode breaks a research question into a multi-step search plan, runs a series of targeted searches, screens a large set of papers and writes a cited report structured like a literature review, with visuals on scientific agreement, key authors, key claims and research gaps. Other tools include Ask Paper, the Consensus Meter, study snapshots, citation export and collections. It runs in the web browser.

Open Code Review is a CLI that reviews Git diffs by combining deterministic pipelines with an LLM agent and posts line-level comments; it works with OpenAI- and Anthropic-compatible model endpoints and can also scan whole files.

Web-based AI design agent that turns a text brief or uploaded brand materials such as logos and brand guides into marketing, e-commerce and branding assets, including product shoots, posters, logos, merchandise designs and video. Work happens on one canvas, where it selects among third-party image and video models (such as GPT Image, Nano Banana, Kling, Veo and Flux) or lets users choose, with editable text layers, a saved brand kit and reusable custom skills.

An agent that takes natural-language instructions and carries out multi-step tasks and test workflows on Android devices or emulators through ADB using multimodal models, with an MCP server so coding assistants can drive devices.

An open-source multi-agent system that collects and analyses social-media posts and comments across platforms and writes public-opinion research reports; built in Python without an external agent framework.

Google's terminal interface (`agy`) to the Antigravity agent platform, built in Go and announced at Google I/O on 2026-05-19 as the replacement for Gemini CLI. It supports background and multi-agent workflows.

Self-hostable workspace where people and AI agents share channels, and messages, workflows, code reviews and Git events are stored as signed events on a Nostr relay the team runs. Agents have their own keys and can open repositories, send patches and run workflows.

Connect your users to their apps

A Python framework for building multi-agent applications with ReAct agents, tool and MCP management, model abstractions, pipelines and human-in-the-loop controls.

Customer service AI agent platform from Forethought, acquired by Zendesk in March 2026 and now marketed as Forethought AI Agents by Zendesk. The Solve agent resolves issues over chat, email, voice, Slack or a headless API; Triage classifies and routes tickets, Discover finds knowledge gaps, QA scores human agents' tickets and Assist suggests replies. Agents learn from past tickets and help-center content and run on Zendesk or other helpdesks such as Salesforce and ServiceNow.

Multica is a project board where issues are assigned to coding agents such as Claude Code, Codex and Cursor alongside human teammates; agents pick up tasks, comment on progress and hand back results for review. It runs as a hosted cloud or self-hosted, with agents on user-controlled runtimes.

Vane is an AI-powered answering engine.

Conversational vibe-coding platform that builds full-stack web and mobile apps with AI agents, with private hosting, GitHub integration and agent building. It offers Free, Standard and Pro individual plans plus Business and Enterprise.

Never stop coding. Free MIT AI gateway: one endpoint, 359 providers (150+ free), 1200+ models Kimi, Claude, GPT, Gemini, GLM, DeepSeek, MiniMax. Works with Claude Code, Codex, Cursor, OpenCode, Cline & Copilot. Quota-aware auto-fallback, RTK+Caveman compression saves 15-95% tokens, MCP/A2A, Desktop/PWA. Built by hundreds of contributors

TencentDB Agent Memory is a team-level memory hub for AI Agents — turning conversations, docs, and code into four reusable memory assets (Chat Memory, Skill, LLM-Wiki, Code-Graph) that are governed, shared, and equipped across agents and frameworks.

Browser-based AI app builder from StackBlitz that generates, runs and deploys full-stack web apps and sites from prompts using in-browser WebContainers. It includes hosting, databases, team plans and Bolt Forge for open-source models.

Sourcegraph's AI coding assistant for VS Code, JetBrains, Visual Studio, the web and a CLI, using code-search context. Free and Pro plans were discontinued in 2025 and it remains only as part of Sourcegraph Enterprise.

xAI's terminal coding agent (`grok`): a full-screen TUI that edits files, runs shell commands and searches the web, with a headless/CI mode and Agent Client Protocol support for editors. Its Rust source is published under Apache-2.0 and synced from an internal monorepo.

A minimal Hugging Face library for agents that write and run Python code as their actions.

CloakBrowser is a Chromium build with fingerprint patches at the C++ source level, used as a drop-in replacement for Playwright and Puppeteer in Python and JavaScript. The wrapper is MIT-licensed; the binary comes in a free version and paid Pro builds.

Obscura is a lightweight headless browser written in Rust for AI agents and web scraping that starts isolated sessions quickly and is driven over the Chrome DevTools Protocol with Playwright or Puppeteer. It can be self-hosted in a Docker container or used as a managed cloud.

Inworld offers realtime speech-to-speech, TTS, STT and LLM routing APIs that game studios can use for NPCs that respond to player speech, world state and story context, with viseme timestamps for lip-sync. It works from Unreal, Unity or custom engines over REST, WebSocket or WebRTC, and games are now one of several markets alongside companions, education and voice agents.

Pixelle-Video generates a short video from a single topic: it writes the script, creates AI images or video clips, synthesizes narration, adds background music and composes the final video. It runs locally and calls LLM, TTS and ComfyUI/RunningHub backends.

AI agents running research on single-GPU nanochat training automatically

Open-source command-line tool (gws) for Drive, Gmail, Calendar, Sheets, Docs, Chat and other Workspace APIs with JSON output and 40+ agent skills; the README says it is not an officially supported Google product.

PicoClaw is an open-source personal AI assistant written in Go that runs as a single low-memory binary, including on low-cost RISC-V boards, and connects chat platforms to LLM providers with tools, MCP servers and web search.

A self-hostable agent that reviews, describes and suggests improvements on pull requests for GitHub, GitLab, Bitbucket, Azure DevOps and Gitea with a configurable model. Qodo handed it to an independent community organisation in 2026.

A self-hosted assistant designed to run on local models that can browse the web, write and run code and split tasks across specialised agents; it runs with Docker and supports optional cloud model APIs.

Dexter is a terminal agent that answers financial research questions by planning research steps, pulling financial statements and market data through APIs, and checking its own results. It is meant for research and education, not for trading.

Distilly — Distill how they think into reusable Skills for any Agent or Bot. Formerly Colleague Skill（原同事 Skill）.

OpenWiki is a CLI that turns a codebase and knowledge sources into a linked Markdown wiki for coding agents, tracking facts back to source evidence; it integrates with Claude Code, Codex, Cursor, OpenCode and other agents and serves wiki content over MCP.

CLI (bsk) and Chrome/Edge extension that let shell-capable AI agents such as Claude Code, Codex and Cursor use the user's logged-in browser in a separate Agent Window to read pages, fill forms, take screenshots and inspect network requests.

cmux is a macOS terminal built on Ghostty with vertical tabs, split panes, an in-app browser and notifications when coding agents such as Claude Code or Codex need attention. It is scriptable through a CLI and socket API.

Open-source memory platform for AI agents that turns documents and conversations into a knowledge graph agents can search and reuse across sessions.

An Open Phone Agent Model & Framework. Unlocking the AI Phone for Everyone Winding down

Open-source Grok Bot alternative. Choose your own model and sandbox.

CodexBar is a macOS menu bar app, with a Linux desktop app and CLI, that shows usage limits, credits, costs and reset times for AI coding providers such as Codex, Claude, Cursor and Gemini, reusing existing provider sessions or API keys.

Toolkit that lets AI agents rebuild a reference video's shots, captions and effects around your own content and render many variants from one command; open core plus a hosted web app and API.

Context Mode is an MCP server and plugin for AI coding agents that runs tool output in a sandbox to keep raw data out of the context window and stores session events in SQLite for retrieval after compaction. A hosted control plane is sold as paid plans.

Moonshot AI's coding agent, offered as the Kimi Code CLI (`kimi`), a desktop app and a VS Code extension; it is included in Kimi memberships and can also use a Moonshot API key or other providers. The CLI was rebuilt in Node.js in 2026 and replaces the archived Python Kimi CLI.

Microsoft's SDK for adding LLMs, plugins and agents to C#, Python and Java applications.

An AI QA agent that explores a running app from a short description and reports findings, plus AI-written regression tests that update themselves when the UI changes and run locally or in CI.

An open-source TypeScript framework for building, orchestrating and observing AI agents.

AI app builder that generates full-stack apps with a built-in backend, database, authentication and hosting from prompts, and also offers a CLI and SDK. It was founded by Maor Shlomo and acquired by Wix.

Open-source desktop office suite (docs, sheets, slides, PDF, Markdown, HTML) with a built-in AI agent, plus a genoffice CLI and agent skill so Claude Code, Codex or Cursor can create and edit real .docx, .xlsx and .pptx files; bring your own model key.

Agent skill and plugin that turns a video URL or local file into timestamped frames and a transcript so Claude Code, Codex and other Agent Skills hosts can analyze it. It uses native captions first and optional WhisperX, Groq or OpenAI transcription.

Ralph is an autonomous AI agent loop that runs repeatedly until all PRD items are complete.

Open-source AI pair-programming tool for the terminal that edits files in a local git repository and commits the changes. It works with multiple LLM providers using the user's own API keys.

Platform for building, deploying and running AI agents through a chat-based AutoPilot, a visual builder and a marketplace, offered as a paid hosted service or self-hosted. The original 2023 autonomous agent remains in the repository as AutoGPT Classic.

ByteDance's open-source visual platform for developing AI agents, released from its Coze product.

Superset is a desktop app for macOS (Linux experimental) that runs CLI coding agents such as Claude Code and Codex in parallel workspaces, with built-in terminals, diff review, PR workflows and browser previews.

Free desktop browser for AI agents that lets coding agents such as Codex or Claude Code run web automation with your logged-in browser state.

Gas Town is a workspace manager that coordinates many AI coding agents such as Claude Code, Codex and Gemini across projects, with a coordinator agent, worker agents and a merge queue. Work state is kept in git-backed hooks and the Beads issue tracker so it survives agent restarts.

Huobao Drama is a self-hosted platform and desktop app that turns a story or novel into a short drama video: built-in agents rewrite the script, extract characters and scenes, break it into storyboards and generate prompts, then image and video models produce clips that are composited with FFmpeg.

Platform for building retrieval-based AI agents from a company's own documents and website content, available as a chat widget and through a RAG API with official SDKs. It targets support and research questions and offers custom enterprise builds.

OpenFang is an open-source agent operating system written in Rust and shipped as a single binary; it runs scheduled autonomous agents ("Hands") with tools, messaging-channel adapters and a local web dashboard.

Platform for building AI support and sales agents trained on a company's own content and published to a website widget, WhatsApp, email, Slack and other channels. Plans are metered in monthly message credits, with model choice across several providers.

Secure, Fast, and Extensible Sandbox runtime for AI agents.

An open-source automation platform with AI agent steps and MCP support, positioned as an alternative to closed workflow tools.

An AI research platform that searches academic papers and runs a research agent, automated research reports and systematic literature review workflows (screening and data extraction). It spun out of the nonprofit Ought.

Source-available 'AI operating system' that runs agents on cloud computers, each session in an isolated sandbox on its own git branch, with agents, skills, memory and connectors kept in one repository and work delivered through reviewable change requests. It works in Slack and Microsoft Teams, with any model provider and your own API keys, and grew out of the Suna agent project.

Opik is an open-source platform for tracing, evaluating and monitoring LLM applications and AI agents, with LLM-as-a-judge metrics, prompt optimization and dashboards; it can be self-hosted or used as Comet's hosted cloud.

ByteDance's AI IDE (branded TraeCode) with an agent mode, offered alongside TraeWork, an AI work assistant with cloud tasks and a web version.

Solution for long term memory for agent coding CLIs and to facilitate handoff between different agent vendors

Bridge that connects locally running coding agents such as Claude Code, Codex, Cursor and Gemini CLI to messaging apps including Slack, Telegram, Discord, Feishu/Lark, DingTalk and WeCom, so users can work with the agent from a chat.

AI customer service agent from Tidio that now has its own site and pay-as-you-go credit pricing, and can be placed on Tidio chat or invited into other help desks such as Zendesk, Freshdesk, Salesforce and HubSpot. It lets customers choose among Claude, GPT and Gemini models.

SciSpace (formerly Typeset) is an AI research workspace whose agent runs multi-step literature reviews, systematic-review screening, data extraction from papers and database searches.

Visual chatbot and chat-form builder that can be embedded on websites or run on WhatsApp and self-hosted. The code is source-available under the Functional Source License (FSL-1.1-ALv2) rather than an OSI-approved open-source license.

An open-source platform for building enterprise agents that combines retrieval pipelines, an agentic workflow engine, MCP tool use and multi-model support, commonly used for customer service and internal knowledge bases.

OpenHarness is an open-source agent harness providing tool use, skills, memory and multi-agent coordination through the oh CLI; it includes ohmo, a personal agent reachable from Feishu, Slack, Telegram and Discord.

Turbocharge Claude Code, Cursor, Codex, Gemini & every coding agent: faster, cheaper, with contextual understanding specific to your codebase.

Local-first desktop workspace for Claude Code and other coding agents: multi-agent teams, Git worktrees, diffs, a skills marketplace, several model providers and remote access from chat apps such as WeChat, Feishu and Telegram.

Odysseus is a self-hosted web workspace for local or API language models that combines chat with autonomous agents and MCP tools, deep research, model comparison, a document editor, an IMAP email assistant, notes, tasks and calendar. It runs via Docker Compose.

An open-source multi-agent framework and research community focused on agent societies and large-scale simulation.

QM is a self-hosted agent harness for teams that runs in Slack and on the web, giving each person and channel its own scoped memory, files, permissions and sandbox. It deploys to the operator's own cloud and can drive Pi, OpenCode, Codex or Claude Code with chosen models.

An open-source multi-agent penetration testing tool that runs dynamic tests against applications, validates findings with working proofs of concept and proposes fixes; it runs locally with Docker or as a hosted service.

Trellis stores specs, tasks and session journals in a repository's .trellis directory and injects them into AI coding agent sessions so agents follow project conventions. It is installed via npm and supports Cursor, Codex, OpenCode, Claude Code and other platforms.

Local memory server for coding agents that records what an agent did and returns relevant context in later sessions. It works with Claude Code, Cursor, Codex, Gemini CLI and other MCP clients and is started with npx.

📡 Your own AI-powered news radar. Generates daily briefings in English & Chinese. | 用 AI 构建你专属的新闻雷达

Cloudflare OS is a self-hostable AI workspace on Cloudflare Workers with an agent chat, sandboxed AI-built apps called gadgets, and a Gatekeepers layer that controls and logs agent access to external services. It is in early access.

Desktop app, formerly Auto Claude, that plans, builds and validates code with multiple autonomous coding-agent sessions tracked on a Kanban board. The 2.x app is in maintenance mode while version 3.0 with cloud features is developed.

ChatGPT thinks. Codex works. Use ChatGPT as the planning brain while keeping the Codex harness.

A small Python agent that gives a model control of a local computer through the browser, terminal, files, keyboard and mouse, screen vision and Android ADB, and saves each completed task as a reusable skill.

Cross-platform local MCP capabilities for ChatGPT with Chrome integration, Goal, Compact & Resume, and durable multi-agent workflows.

Paseo is an open-source agentic development environment that runs Claude Code, Codex, Copilot, OpenCode, Pi and other coding agents in parallel worktrees on the user's machines, with editor, terminals, diffs and a browser across desktop, iOS, Android, web and CLI.

Microsoft's open-source SDK for building, orchestrating and deploying agents and multi-agent workflows in Python and .NET. It is the successor to both AutoGen and Semantic Kernel.

Reviews GitHub and GitLab pull requests with bug findings, summaries, diagrams, one-click fixes and security scanning, and offers in-IDE reviews.

MemPalace stores conversation history verbatim and retrieves it with semantic search over an index organized into wings, rooms, closets and drawers. It runs locally as a Python CLI and MCP server with a pluggable backend, ChromaDB by default.

Memory plugin built for Claude Code: it captures what the agent does during a session, compresses it into AI-written summaries and injects the relevant context into later sessions, so work continues after a session ends. It also installs for Codex, Gemini, OpenCode, Copilot, OpenClaw and other agents. The engine runs locally for free; CMEM Pro adds a hosted observer, cloud sync across machines and a private MCP link.

Runs codebase-aware AI reviews on pull requests in GitHub, GitLab and Bitbucket and in the IDE, with an optional self-hosted Docker agent.

An AI code review platform that reviews pull requests across repositories using codebase context and team rules, with IDE extensions for local review.

Creates, runs and repairs tests for web, mobile, API, accessibility and performance, and describes itself as an agentic testing platform.

A self-hosted multi-agent system that carries out penetration testing tasks in a sandboxed environment, written in Go and React and working with several model providers.

An open-source, privacy-first, self-hosted knowledge workspace where humans and AI agents work together 开源、隐私优先、自托管的知识工作空间，让人与智能体在此协作

Open-source agentic workspace enterprises can make their own. Connect the systems you already run — 100+ integrations, MCP, chat tools, apps, browser, local files — with shared memory. Any agent (Claude Code, Codex), any model, or BYOK. Set up in clicks, not months. Local-first: your data never leaves your machines.

Godot MCP is an open-source MCP server that lets AI assistants launch the Godot editor, run projects in debug mode, capture debug output and inspect project structure. It can also create scenes, add nodes, load sprites, export MeshLibrary resources and manage UIDs (Godot 4.4+).

A data notebook and analytics platform whose Notebook Agent writes and runs SQL and Python to investigate questions, while Threads answers natural-language data questions and a modelling agent builds semantic models. Context Studio shows the context used, queries run and steps taken for each answer (as described in Hex's docs).

Open-source multi-agent harness for running AI agents in production.

An AI scientist agent from Edison Scientific, a commercial spinout of FutureHouse. A run reads large amounts of literature, executes analysis code and produces cited reports.

Open-source headless browser for AI agents, written in Rust and built to be light and fast for agent web automation.

Mistral's coding agent, used as the `vibe` CLI, a VS Code extension (with a JetBrains plugin and web version) and powered by Devstral models. Since May 2026 Mistral brands its whole assistant, formerly Le Chat, as Vibe.

AI support agent from Crisp that runs on a Crisp workspace. Hugo answers customer conversations from a company's knowledge base, docs, FAQs and CRM data, keeps context across multi-turn chats, resolves some requests end to end through tool access and hands off to a human with context when needed. Workflows are built in a no-code visual editor; it can use Claude, ChatGPT, Llama or a custom model, connects to tools over MCP and processes data on European servers.

Enterprise customer experience AI platform from Netomi (formerly msg.ai) with a no-code orchestration layer that coordinates a system of AI agents. The agents handle customer requests over chat, email and voice, from routine questions to multi-step workflows, and leave complex cases to human staff, with governance and brand-compliance controls. Accenture Ventures led a 2026 investment and the two companies formed a strategic partnership. New engagements are by referral.

PinchTab is a standalone Go HTTP server that gives AI agents control of Chrome instances through an HTTP API, MCP server and CLI, typically run as a local daemon with a dashboard.

Open-source desktop app for planning and orchestrating AI coding agents (Claude Code, Codex, Cursor, OpenCode and others) in parallel, each in its own git worktree. Turns ideas into specs and tickets, supports agent-to-agent communication, code review of diffs, shared context across models, and team collaboration.

A Python library that turns natural-language questions into Python and SQL, runs it against dataframes, databases and data lakes, and returns answers and charts.

Experiential is the open source, zero markup gateway for BYOK, self-hosted and 1000+ marketplace models. It learns from your traffic to cut costs, recommend better models, and train a specialized model you own.

AWS service and Apache-2.0 Python SDK for building and running fleets of UI-automation agents that complete browser workflows from natural-language act() calls plus Python code and hand off to a human when needed. Workflows are developed in a web playground, IDE extension or CLI and deployed on AWS.

Praxist runs persistent research loops on projects with a measurable objective, coordinating parallel research agents, task-owned evaluation, evidence retention and synthesis across generations. It is installed via pip and operated through Codex or Claude Code.

Supermemory is a memory and context layer for AI applications that extracts facts from conversations, maintains user profiles and combines memory with RAG search. It is used through an API, an MCP server or plugins for coding agents, hosted or self-hosted.

Posts automated reviews, summaries and suggested fixes on pull requests, and also runs reviews in the IDE and from a command-line tool.

An open-source agent that plans research, gathers and filters many web or local sources and writes a cited report, with support for multiple model providers.

Open-source operating system for AI agents that provides the aos CLI, an HTTP API and WebAssembly capsules running on the Astrid runtime, with capability tokens and an audit log recording agent actions.

OpenMuse is a self-hosted personal agent app built with CopilotKit and AG-UI that works in a persistent browser, an optional Linux terminal and files, with task plans, approvals and Gmail/Calendar integration on iOS, Android and web.

AWS's terminal agent from the Kiro family (`kiro-cli`) for building applications with natural-language chat and automated workflows; sign-in uses Kiro accounts (social login or AWS Builder ID) and plans are credit-based.

video-use is a skill for Claude Code, Codex and other shell-capable agents that edits raw footage into a finished video through chat: it cuts filler words, color grades, adds subtitles and generates animation overlays using ffmpeg and ElevenLabs transcription.

UI design tool from Uizard Technologies whose Autodesigner generates multi-screen, editable app and website prototypes from text prompts, matched to a brand kit. Users refine results by selecting a component and describing the change, or by prompting for new screens and themes. The app also has Screenshot and Wireframe Scanners that turn screenshots and sketches into editable mockups, plus templates and real-time collaboration. Uizard joined Miro in June 2024.

A provider-neutral terminal coding agent written in Rust that reads projects, edits files, runs commands and tests to check its work, with approval-based permissions, undo and sub-agents. It grew out of the deepseek-tui project.

Platform for building AI agents and assistants for customer experience and documentation, with a visual builder, a TypeScript SDK and chat components, offered as a source-available framework under the Elastic License 2.0 plus a managed Enterprise edition. Enterprise adds Slack, Discord, Teams, Zendesk and Salesforce integrations.

Desktop app that runs several coding agents in parallel, giving each task its own branch and worktree and tracking pull requests, CI and reviews on a Kanban board. A project-level orchestrator agent plans work and delegates tasks to worker agents.

A self-hosted assistant, formerly chatgpt-on-wechat, that plans tasks and runs tools, skills and browser automation with memory, a knowledge base and multi-agent teams, connecting to web, WeChat, Feishu, DingTalk, Telegram, Slack and other channels.

The open-source agent harness - the runtime layer that turns an LLM into a working agent.

AI customer agent that answers and resolves support conversations over chat, email, voice and messaging channels, either natively inside the Intercom helpdesk or on top of other help desks such as Salesforce, HubSpot and Freshdesk. It runs on the vendor's own Apex model suite. Salesforce reported completing its acquisition of Fin on September 10, 2026.

A community-driven continuation of the original AutoGen codebase, maintained as an open-source multi-agent framework.

AI-powered virtual executive team — a single coherent executive persona backed by 8 specialist agents (FastAPI + Next.js).

Coding agent and development environment from Amp, spun out of Sourcegraph, used through a CLI, a web app and macOS/iOS apps; each thread can run in a cloud 'orb' that keeps working after the laptop is closed. Users can bring their own model keys or link a ChatGPT subscription.

RTK is a single Rust binary that intercepts shell commands run by coding agents and compresses their output (git, ls, grep, test runners and 100+ other commands) before it reaches the LLM context, reducing input tokens.

Charm's terminal-based AI coding agent (TUI) that connects to many LLMs through OpenAI- and Anthropic-compatible APIs and supports LSP and MCP. Its source is public under the FSL-1.1-MIT license, which is source-available rather than OSI open source.

Vercel's AI builder that turns prompts into full-stack web apps and UIs, with GitHub sync (branches and pull requests), Vercel deployment, a visual Design Mode, a model picker and an iOS app.

Platform for designing and deploying AI agents that run across web, app, WhatsApp, SMS and voice, aimed at agencies building for clients and at businesses running customer experience. It offers a model picker across major providers and a Shopify integration app.

Prime Agent is a terminal coding and research agent built around a Recursive Language Model harness: the model works through a persistent Python REPL, spawns subagents and can refine its own prompts, memories and skills. Sessions can run in the background and be reattached.

Open-source personal trading agent from HKUDS that researches markets and builds and backtests trading strategies from plain-language requests.

Skill, local proxy and SDK that reduce the tokens coding agents write and read: the skill makes replies terse and the proxy compresses logs, JSON, test output and web pages before they reach the model. It works with Claude Code, Codex, Gemini CLI and other agents.

Service for building AI support agents and chatbots trained on a company's websites, files, knowledge bases, cloud storage and closed help desk tickets. Bots answer customer, pre-sales and internal questions, hand off to humans, and can run actions such as Stripe billing, booking and MCP tools. They deploy as a website widget, via API, in help desks such as Zendesk and Help Scout, and on Slack, WhatsApp, Telegram and Messenger. Usage is metered in AI credits.

Open-source retrieval-augmented generation (RAG) engine from InfiniFlow with agent features: it extracts knowledge from complex documents through deep document understanding, keeps it as a searchable knowledge base and lets you build agents on top from pre-built templates, including agentic RAG with several thinking modes. Self-host it with Docker or use the hosted RAGFlow Cloud.

A pull-request review platform for stacked PRs whose AI reviewer posts codebase-aware feedback, explains changes, proposes fixes and helps resolve CI failures. Cursor agreed to acquire Graphite in December 2025; it continues as a separate product.