Daily · 16 June 2026

Top 100 AI Agents of 2026

Ranked from 100 down to 1. Generated by /lad, illustrated by /iad.

Explore Diagram Explore as app Share

Anthropic's terminal-native agentic-coding CLI — the canonical reference implementation for long-horizon coding agents. Hooks, MCP, subagents, plugins. The dev-agent benchmark of 2026.

Cognition's autonomous software engineer — assign a Linear ticket, get a PR. The first AI agent priced like a junior employee ($500/mo) and the original 'AI software engineer' product.

Operator (OpenAI)

OpenAI's browser-controlling agent — runs as a virtual browser, books flights, fills forms, navigates UIs. The flagship consumer-facing agent of the ChatGPT family.

Cursor Background Agents

Cursor's cloud-side autonomous agents — they keep working on your PR even when your laptop is closed. Reshaped what 'IDE' means in 2025-26.

Chinese general-purpose agent that went viral in early 2025 for closing loops on real-world tasks (booking, research, spreadsheets) better than most US rivals.

Anthropic Computer Use

Claude's computer-use API — agents that control a mouse + keyboard inside a VM. The future of OS-level agents.

OpenAI's general agent mode in ChatGPT — combines browsing, coding, file operations. The mass-market on-ramp to agent workflows.

Gemini Deep Research

Google Gemini's multi-step research agent — runs queries, synthesises sources, produces reports. The default 'research agent' for the Google-tied user.

No-code AI agent builder — connects to Gmail, Calendar, Slack, HubSpot. The Zapier-for-agents that found product-market fit in SMBs.

Open-source autonomous coding agent that runs in your IDE — does file edits, runs terminal commands, browses the web. The OSS Devin.

Fork of Cline with deeper customisation and multi-mode personas (Code, Architect, Ask, Debug). The choice when Cline isn't opinionated enough.

OpenHands (OpenDevin)

Open-source Devin alternative — runs a sandboxed agent that edits files, runs commands. Strong academic backing and the most-forked OSS coding agent.

Block (Square) open-sourced agent runner — MCP-native, plugin-rich. The pragmatic OSS choice for shipping agents in production.

Junior dev as a GitHub bot — assign it a ticket, get a PR back. Quietly profitable, niche but well-loved.

AI code review as a GitHub bot — catches what the human reviewer misses. Specialised but accurate.

AI PR reviewer — line-by-line comments, summaries, security checks. Strong adoption across mid-market SaaS.

AI code review + repo-wide code search. Found product-market fit in 2024 after a pivot from agent to review.

AI PR reviewer specifically tuned for AI-generated code — catches hallucinated API calls. Niche but timely.

AI agent that fixes JIRA tickets — generates a PR, runs tests, iterates. Smaller-shop Devin alternative.

Intercom's AI agent that resolves a verifiable 50%+ of support tickets autonomously. The benchmark every AI support tool is judged against.

Zendesk AI Agents

Enterprise support incumbent with deeply-integrated AI agents and copilots. The choice when scale and compliance trump shiny new tools.

Vertical AI customer-service agent — Series C unicorn focused on conversational support quality. Strong in fintech and consumer brands.

Bret Taylor's customer-experience AI agent platform — multi-modal, voice + chat. Series A at $4.5B; the most-funded support-agent startup.

European voice-AI agent for contact centres. The serious enterprise competitor in regulated markets (telco, finance).

Real-time agent assist + autonomous AI for contact centres. Deep into the customer-success enterprise wedge.

AI sales rep that prospects, writes outbound, books meetings. The flagship 'digital worker' marketing concept.

AI BDR named Ava — outbound prospecting at scale. Their 'Stop Hiring Humans' billboards defined the 2024 SF aesthetic.

Outbound sales platform that uses AI to enrich, score, and personalise prospects. The fastest-growing GTM tool of 2025 by far.

Apollo AI Agent

Apollo's outbound AI agent layered on top of their sales-intelligence database. Built into 100K+ sales-team workflows.

Outreach's autonomous sales workflows — sequences, follow-ups, dialler. The enterprise incumbent gone AI-native.

Salesforce Agentforce

Salesforce's agent platform — Atlas reasoning engine, deep CRM data integration. The biggest enterprise bet on agents.

HubSpot's AI agent suite — Customer agent, Content agent, Prospecting agent. SMB-first by design.

Microsoft Copilot Studio

Microsoft's enterprise platform for building custom agents — deep M365 + Dynamics integration. The default for Microsoft shops.

Google Vertex AI Agent Builder

Google Cloud's agent platform — strong on grounding, RAG, and enterprise auth. The Google-shop equivalent of Copilot Studio.

AWS Bedrock Agents

AWS's managed-agent platform on top of Bedrock. Strong story for regulated industries on AWS.

Domain-specific assistant for big law — drafting, research, contract review. The most-funded vertical AI of 2024/25.

Plaintiff-firm AI agent — document review, discovery, deposition prep. Closed-vertical, deeply specialised.

AI contract drafting + review for transactional lawyers. The Word-plugin AI that big firms quietly adopted.

Healthcare voice-AI agent — patient outreach, post-discharge follow-up, nurse-line triage. Voice with safety-engineering at the core.

Clinical AI scribe — records the patient conversation, produces the note. The Epic-integrated category leader.

Voice-first clinical AI assistant — ambient documentation for clinicians. Competitor to Abridge on EHR integration.

Ambience Healthcare

Documentation + medical-coding AI agent across the visit workflow. Series C; deep payer + provider partnerships.

Enterprise search + AI assistant with agent capabilities baked in. The Series F darling of enterprise AI; every Fortune 1000 evaluating it.

Writer AI Agents

Enterprise content + workflow platform with custom-tuned models. Quiet incumbent in the F500 communications stack.

AI meeting notes that records, transcribes, and summarises right inside macOS — no bots in the call. The operator-favourite of 2026.

AI meeting assistant with the broadest integration surface — joins calls, transcribes, tags speakers, pushes to CRM.

Otter's agent that joins meetings autonomously and follows up. The pre-LLM-era product that adapted well.

Empathic Voice Interface — voice agents that read emotion and respond with prosody. The leading specialised voice-AI provider.

Voice-AI agent infrastructure — buy a phone number, give it a system prompt, get a voice agent. The Twilio of voice agents.

Phone-call AI agent platform — outbound + inbound voice automation. Sketchy use cases aside, technically among the best.

Voice-AI agent platform with strong latency story. Builder-developer favourite for voice agents.

Eleven Labs spinoff voice-agent product — combining ElevenLabs voices with conversational AI. The end-to-end pipeline.

On-screen real-time AI assistant — sees your screen and whispers answers (interviews, sales calls). The most-controversial productivity tool of 2025.

AI-native product-management copilot — write PRDs, generate user stories, sync to Linear/Jira. PMs' answer to Cursor.

Lovable's autonomous full-stack app-build agent — chat your way to a deployed Next.js app. The non-dev's Devin.

StackBlitz's in-browser Next.js/Vite generator agent. Won the no-code-to-real-code race in 2025.

Text-to-UI generation agent — outputs production React + Tailwind. The fastest path from spec to shipped frontend.

Cloud IDE plus an autonomous coding agent that scaffolds, deploys, and iterates. The lowest-friction idea-to-URL path.

Open-source library that drives Playwright via LLM tool-calls — the default underpinning for any 'browse-the-web for me' agent.

Stagehand by Browserbase

AI-friendly browser-automation framework — sits on top of Playwright with LLM-native APIs. Cleaner DX than browser-use.

Headless-browser infrastructure for AI agents — sandboxed, fingerprint-managed, captcha-aware. The Vercel of agent browsers.

Open-source LLM-powered browser-automation agent. The OSS Browser-Use alternative with stronger anti-bot resilience.

Browser agent that takes natural-language instructions and executes them across the open web. Consumer-targeted, demo-rich.

The original autonomous-agent meme — recursive goal pursuit, self-prompting. By 2026 a historical artefact but still a teaching tool.

Yohei Nakajima's task-creating agent that catalysed the autonomous-agent imagination in 2023. Tiny code, huge influence.

Open-source multi-agent framework that role-plays product manager + engineer + reviewer. Research-favourite for studying collaboration.

Multi-agent orchestration framework — define agents as crews with roles + tasks. Strong PyPI download numbers and dev mindshare.

AutoGen (Microsoft)

Microsoft Research's multi-agent conversation framework. The reference implementation for multi-agent research patterns.

LangChain's graph-based agent framework — state machines for agents. The choice when CrewAI/AutoGen are too rigid.

TypeScript-native agent framework with built-in observability, memory, and workflow primitives. The 2026 favourite for production TS agents.

Pydantic-first Python agent framework — typed in, typed out. Strong DX for FastAPI-shop teams.

Docs-search-as-a-service agent for product teams. The post-Algolia bet on conversational docs.

Anthropic Claude Agents API

Anthropic's first-party agentic primitives — tool use, computer use, MCP. The closest a frontier lab gets to giving you the actual agent kit.

OpenAI Assistants API

OpenAI's stateful assistants with code interpreter, retrieval, function calling. The oldest production agent platform still standing.

Mistral Agents API

Mistral's structured-output + function-calling agentic layer. EU-data-resident option for compliance-sensitive teams.

ServiceNow Now Assist

ServiceNow's AI agents inside ITSM workflows. The enterprise heavyweight you don't see on Twitter.

Notion AI Agents

AI baked into the most-used productivity workspace — Q&A on docs, drafting, summarisation, agentic actions across the workspace.

Linear's AI suggestions across issue triage and project routing. The PM tool whose AI features stay invisible until you need them.

Atlassian's AI teammate across Jira + Confluence — context-aware across the wiki + tickets. Big enterprise rollout.

Asana AI Studio

Asana's AI-builder for custom workflow agents inside the work-graph. The latest entry in PM-AI.

Monday's AI assistant + agent builder. SMB-first work-management AI.

Zapier's agent platform on top of their integration graph — the 'agent that can do anything' for the Zapier user base.

Open-source workflow automation with strong AI-agent node support. Self-hostable, popular with technical teams.

Make's no-code agent layer on top of their visual workflow builder. The Zapier alt for AI-first SMB teams.

Visual workflow builder for AI pipelines. The choice when Zapier is too rigid and code is too much.

AI agent platform with strong sales + support team templates. Australian-founded, strongest in APAC and mid-market.

Visual agent-builder targeted at enterprises — RBAC, SOC 2, on-prem options. Niche but profitable.

Observability + evaluation platform purpose-built for agentic apps. Loved by teams running production agents.

Open-source LLM-observability platform with strong agent-trace support. The OSS alt to LangSmith for agentic apps.

LangChain's tracing + evals platform. Standard for any team running agents in prod across the LangChain ecosystem.

Letta (formerly MemGPT)

Long-term-memory agent framework — agents that learn across sessions. Research-heavy, open-source.

Memory layer for AI agents — embeds + ranks past interactions across sessions. The de-facto choice when LangGraph's checkpointer isn't enough.

Pieces for Developers

Persistent code-snippet + context manager with AI agent search across dev history. The dev-knowledge-graph play.

Multi-agent coding research lab; Tristan is their public agent demo. Niche but technically respected.

Frontier-lab focused on autonomous coding agents with extremely long context. Series A at $1.5B valuation, mostly a research bet.

Pioneer of action-transformer agent research, acquired by Amazon in 2024. Foundational influence on browser-agent thinking.

Pythagora / GPT Pilot

Open-source AI dev that asks clarifying questions before writing code. The 'good engineer' archetype agent.

Open-source AI pair programmer in the terminal — git-aware, multi-file edits, model-agnostic. The hacker-crowd's favourite.

Open-source terminal coding agent; lightweight, model-agnostic. Beloved by indie hackers who want minimal surface area.

Goose Plugins (MCP)

Anthropic's Model Context Protocol — the open standard for connecting agents to tools. Adopted by Cursor, Claude Code, Goose, Zed.

01

Lv 1 · Browser0 pts

0 / 100 to Lv 2+1 / 200px scrolled

Theme

Display

Density