General-purpose agents
Best forKnowledge workers delegating substantial multi-step research, analysis, browser and file-based work.
Not ideal forSimple one-shot chat tasks or teams that require every action to remain purely manual.
THE AGENT INTELLIGENCE DIRECTORY
AI Agents
& Models
Find the right offering for the work. Compare what it’s best for, where it falls short and how it fits into your workflow.
More filters
| Offering / vendor | Where it fits | Price / access | Channel / autonomy | Actions |
|---|---|---|---|---|
OpenAI General-purpose agents | Best for Knowledge workers delegating substantial multi-step research, analysis, browser and file-based work. Not ideal forSimple one-shot chat tasks or teams that require every action to remain purely manual. | Paid/plan-dependent Starting price: check provider | Web + mobile + desktop + cloud browser High autonomy | Visit site |
AT A GLANCE ChatGPT WorkOpenAI’s current general-purpose agentic workspace for long-running, multi-step work across files, apps and the web.
PRACTICAL FIT Pricing and permissions can change. Confirm plan limits and approval controls with the provider. Evaluation tracks group comparable work; they are not performance rankings. | ||||
 Anthropic General-purpose agents | Best for Knowledge workers, developers and teams needing strong reasoning, research and coding. Not ideal forTeams wanting a tightly deterministic workflow engine or deeply app-specific automation. | Paid + free access Starting price: check provider | Web + desktop Medium–High autonomy | Visit site |
AT A GLANCE ClaudeOne of the leading general assistants with strong research, coding and computer-use capabilities.
PRACTICAL FIT General-purpose agentsBest forKnowledge workers, developers and teams needing strong reasoning, research and coding. Not ideal forTeams wanting a tightly deterministic workflow engine or deeply app-specific automation. Pricing and permissions can change. Confirm plan limits and approval controls with the provider. Evaluation tracks group comparable work; they are not performance rankings. | ||||
 Google General-purpose agents | Best for Google Workspace users and teams combining research with Google ecosystem workflows. Not ideal forTeams outside the Google ecosystem that need highly customised cross-app automation. | Paid + free access Starting price: check provider | Web + Android + Google ecosystem Medium–High autonomy | Visit site |
AT A GLANCE GeminiGoogle-distributed assistant with search, Workspace and computer-use capabilities.
PRACTICAL FIT General-purpose agentsBest forGoogle Workspace users and teams combining research with Google ecosystem workflows. Not ideal forTeams outside the Google ecosystem that need highly customised cross-app automation. Pricing and permissions can change. Confirm plan limits and approval controls with the provider. Evaluation tracks group comparable work; they are not performance rankings. | ||||
Microsoft General-purpose agents | Best for Microsoft 365 and Windows users who want AI embedded in everyday work. Not ideal forTeams that do not use Microsoft’s ecosystem or need a developer-first autonomous agent. | Paid + free access Starting price: check provider | Web + Windows + Microsoft 365 Medium–High autonomy | Visit site |
AT A GLANCE Microsoft CopilotMass distribution through Windows and Microsoft 365; important enterprise baseline.
PRACTICAL FIT General-purpose agentsBest forMicrosoft 365 and Windows users who want AI embedded in everyday work. Not ideal forTeams that do not use Microsoft’s ecosystem or need a developer-first autonomous agent. Pricing and permissions can change. Confirm plan limits and approval controls with the provider. Evaluation tracks group comparable work; they are not performance rankings. | ||||
 Perplexity General-purpose agents | Best for Fast web research, source-led answers and current-information discovery. Not ideal forDeterministic operational workflows or complex multi-app task execution. | Paid + free access Starting price: check provider | Web + apps Medium–High autonomy | Visit site |
AT A GLANCE PerplexityStrong research and answer-engine brand with agentic browsing/research surfaces.
PRACTICAL FIT General-purpose agentsBest forFast web research, source-led answers and current-information discovery. Not ideal forDeterministic operational workflows or complex multi-app task execution. Pricing and permissions can change. Confirm plan limits and approval controls with the provider. Evaluation tracks group comparable work; they are not performance rankings. | ||||
Manus General-purpose agents | Best for Founders and operators delegating multi-step web research and execution. Not ideal forHighly controlled enterprise workflows requiring strict deterministic behaviour and governance. | Paid / invite-plan dependent Starting price: check provider | Web / cloud High autonomy | Visit site |
AT A GLANCE ManusHigh-profile autonomous-agent product built around completing multi-step tasks.
PRACTICAL FIT General-purpose agentsBest forFounders and operators delegating multi-step web research and execution. Not ideal forHighly controlled enterprise workflows requiring strict deterministic behaviour and governance. Research agentsBest forFounders and analysts combining research with execution and finished deliverables. Not ideal forHighly governed research environments requiring strict deterministic controls. Pricing and permissions can change. Confirm plan limits and approval controls with the provider. Evaluation tracks group comparable work; they are not performance rankings. | ||||
Genspark General-purpose agents | Best for Users combining web research, content creation and multi-step task execution. Not ideal forTeams needing deep enterprise administration or highly specialised vertical workflows. | Freemium/paid Starting price: check provider | Web / cloud High autonomy | Visit site |
AT A GLANCE Genspark Super AgentAgent-first product spanning research, creation and execution; useful challenger to Manus.
PRACTICAL FIT General-purpose agentsBest forUsers combining web research, content creation and multi-step task execution. Not ideal forTeams needing deep enterprise administration or highly specialised vertical workflows. Pricing and permissions can change. Confirm plan limits and approval controls with the provider. Evaluation tracks group comparable work; they are not performance rankings. | ||||
Fellou General-purpose agents | Best for Power users who want an agentic browser for multi-step web tasks. Not ideal forTeams seeking mature enterprise governance or a broad non-browser automation platform. | Freemium/paid Starting price: check provider | Desktop/browser High autonomy | Visit site |
AT A GLANCE FellouAgentic browser focused on executing multi-step web work; useful browser-agent benchmark candidate.
PRACTICAL FIT General-purpose agentsBest forPower users who want an agentic browser for multi-step web tasks. Not ideal forTeams seeking mature enterprise governance or a broad non-browser automation platform. Pricing and permissions can change. Confirm plan limits and approval controls with the provider. Evaluation tracks group comparable work; they are not performance rankings. | ||||
Anthropic Coding agents | Best for Developers and engineering teams delegating repo-scale coding and terminal work. Not ideal forNon-technical users or teams wanting a visual no-code experience. | Paid plan/API Starting price: check provider | CLI + IDE + web + mobile High autonomy | Visit site |
AT A GLANCE Claude CodeCategory-leading coding agent for repo-scale autonomous work.
PRACTICAL FIT Coding agentsBest forDevelopers and engineering teams delegating repo-scale coding and terminal work. Not ideal forNon-technical users or teams wanting a visual no-code experience. Pricing and permissions can change. Confirm plan limits and approval controls with the provider. Evaluation tracks group comparable work; they are not performance rankings. | ||||
OpenAI Coding agents | Best for Developers delegating substantial coding tasks across cloud, CLI and IDE workflows. Not ideal forNon-technical teams or simple one-off code-completion needs. | Paid plan/API Starting price: check provider | Cloud + CLI + IDE High autonomy | Visit site |
AT A GLANCE OpenAI CodexMajor coding-agent platform with broad ChatGPT/OpenAI distribution.
PRACTICAL FIT Coding agentsBest forDevelopers delegating substantial coding tasks across cloud, CLI and IDE workflows. Not ideal forNon-technical teams or simple one-off code-completion needs. Pricing and permissions can change. Confirm plan limits and approval controls with the provider. Evaluation tracks group comparable work; they are not performance rankings. | ||||
GitHub / Microsoft Coding agents | Best for Teams already centred on GitHub that want issues turned into reviewed pull requests. Not ideal forRepositories outside GitHub or teams wanting a standalone AI-native IDE. | Paid Starting price: check provider | GitHub + IDE High autonomy | Visit site |
AT A GLANCE GitHub Copilot Coding AgentDeep GitHub integration and enterprise distribution make it a must-benchmark product.
PRACTICAL FIT Coding agentsBest forTeams already centred on GitHub that want issues turned into reviewed pull requests. Not ideal forRepositories outside GitHub or teams wanting a standalone AI-native IDE. Pricing and permissions can change. Confirm plan limits and approval controls with the provider. Evaluation tracks group comparable work; they are not performance rankings. | ||||
Anysphere Coding agents | Best for Developers wanting an AI-native IDE for fast interactive and agentic coding. Not ideal forTeams committed to another IDE or requiring a primarily asynchronous delegation model. | Freemium/paid Starting price: check provider | Desktop IDE + cloud High autonomy | Visit site |
AT A GLANCE CursorOne of the most visible AI-native coding environments and agent surfaces.
PRACTICAL FIT Coding agentsBest forDevelopers wanting an AI-native IDE for fast interactive and agentic coding. Not ideal forTeams committed to another IDE or requiring a primarily asynchronous delegation model. Pricing and permissions can change. Confirm plan limits and approval controls with the provider. Evaluation tracks group comparable work; they are not performance rankings. | ||||
Cognition Coding agents | Best for Engineering teams delegating complete, asynchronous software-development tasks. Not ideal forDevelopers who mainly want fast inline assistance and tight hands-on control. | Paid Starting price: check provider | Cloud Very high autonomy | Visit site |
AT A GLANCE DevinDistinct delegation model: users hand off complete engineering tasks rather than co-editing.
PRACTICAL FIT Coding agentsBest forEngineering teams delegating complete, asynchronous software-development tasks. Not ideal forDevelopers who mainly want fast inline assistance and tight hands-on control. Pricing and permissions can change. Confirm plan limits and approval controls with the provider. Evaluation tracks group comparable work; they are not performance rankings. | ||||
Cognition Coding agents | Best for Developers wanting an IDE plus parallel local/cloud agent workflows. Not ideal forTeams wanting a terminal-only agent or those unable to adopt a new desktop coding environment. | Freemium/paid Starting price: check provider | Desktop IDE + local/cloud agents High autonomy | Visit site |
AT A GLANCE Devin DesktopCurrent successor name for Windsurf: an IDE plus multi-agent command centre for delegating coding work.
PRACTICAL FIT Coding agentsBest forDevelopers wanting an IDE plus parallel local/cloud agent workflows. Not ideal forTeams wanting a terminal-only agent or those unable to adopt a new desktop coding environment. Pricing and permissions can change. Confirm plan limits and approval controls with the provider. Evaluation tracks group comparable work; they are not performance rankings. | ||||
Replit Coding agents | Best for Founders and non-technical builders turning ideas into working, deployed applications. Not ideal forComplex enterprise codebases needing highly customised local development environments. | Freemium/paid Starting price: check provider | Web IDE + cloud High autonomy | Visit site |
AT A GLANCE Replit AgentStrong fit for ND AI audience because it turns prompts into deployed applications.
PRACTICAL FIT Coding agentsBest forFounders and non-technical builders turning ideas into working, deployed applications. Not ideal forComplex enterprise codebases needing highly customised local development environments. Pricing and permissions can change. Confirm plan limits and approval controls with the provider. Evaluation tracks group comparable work; they are not performance rankings. | ||||
Google Coding agents | Best for Developers delegating asynchronous GitHub coding tasks within Google’s ecosystem. Not ideal forTeams needing a local-first or deeply interactive IDE coding experience. | Access varies Starting price: check provider | Cloud + GitHub High autonomy | Visit site |
AT A GLANCE JulesGoogle-backed asynchronous coding agent; strategically important foundation-lab entrant.
PRACTICAL FIT Coding agentsBest forDevelopers delegating asynchronous GitHub coding tasks within Google’s ecosystem. Not ideal forTeams needing a local-first or deeply interactive IDE coding experience. Pricing and permissions can change. Confirm plan limits and approval controls with the provider. Evaluation tracks group comparable work; they are not performance rankings. | ||||
 AWS Coding agents | Best for AWS-centric development teams that want coding help tied closely to cloud workflows. Not ideal forTeams with little AWS usage or those wanting a vendor-neutral AI-native IDE. | Free + paid Starting price: check provider | IDE + AWS Medium–High autonomy | Visit site |
AT A GLANCE Amazon Q DeveloperImportant enterprise coding agent for AWS-centric teams.
PRACTICAL FIT Coding agentsBest forAWS-centric development teams that want coding help tied closely to cloud workflows. Not ideal forTeams with little AWS usage or those wanting a vendor-neutral AI-native IDE. Pricing and permissions can change. Confirm plan limits and approval controls with the provider. Evaluation tracks group comparable work; they are not performance rankings. | ||||
JetBrains Coding agents | Best for Professional developers already working heavily in JetBrains IDEs. Not ideal forTeams centred on VS Code, browser IDEs or non-JetBrains development workflows. | Paid/plan-dependent Starting price: check provider | JetBrains IDEs High autonomy | Visit site |
AT A GLANCE JetBrains JunieDirect access to JetBrains' large IDE user base.
PRACTICAL FIT Coding agentsBest forProfessional developers already working heavily in JetBrains IDEs. Not ideal forTeams centred on VS Code, browser IDEs or non-JetBrains development workflows. Pricing and permissions can change. Confirm plan limits and approval controls with the provider. Evaluation tracks group comparable work; they are not performance rankings. | ||||
Tabnine Coding agents | Best for Enterprise engineering teams prioritising governance, privacy and controlled AI coding. Not ideal forTeams prioritising the most autonomous end-to-end software-engineering delegation. | Paid + trial Starting price: check provider | IDE Medium–High autonomy | Visit site |
AT A GLANCE TabnineLong-standing enterprise AI coding vendor with governance/security positioning.
PRACTICAL FIT Coding agentsBest forEnterprise engineering teams prioritising governance, privacy and controlled AI coding. Not ideal forTeams prioritising the most autonomous end-to-end software-engineering delegation. Pricing and permissions can change. Confirm plan limits and approval controls with the provider. Evaluation tracks group comparable work; they are not performance rankings. | ||||
Cline Coding agents | Best for VS Code developers wanting an open, model-flexible agent with strong tool use. Not ideal forNon-technical users or teams wanting a managed enterprise coding platform. | Open source / BYOK Starting price: check provider | VS Code High autonomy | Visit site |
AT A GLANCE ClinePopular open-source coding agent with model flexibility and tool use.
PRACTICAL FIT Coding agentsBest forVS Code developers wanting an open, model-flexible agent with strong tool use. Not ideal forNon-technical users or teams wanting a managed enterprise coding platform. Pricing and permissions can change. Confirm plan limits and approval controls with the provider. Evaluation tracks group comparable work; they are not performance rankings. | ||||
Showing 20 of 128
USE THE DIRECTORY WELL
Compare the fit, not the hype
Compare like with like across agents, builders, frameworks and evaluation tools. Evaluation tracks group products intended for similar work.
DIRECTORY NOTESHow to read the research+
Coverage priorities are not test scores and no hands-on benchmark results are claimed. Research dates appear where supplied; overlapping entries share one product record.
Confirm current pricing, availability and permissions on the provider’s website. Retired offerings stay hidden unless you enable “Include retired”.
Free subscriber bonus
Get the AI Monetisation
Library, free when you join
Discover 30 practical ways to make money with AI, from simple service businesses to scalable digital products. Each opportunity includes how it works, the tools to use and the monetisation model behind it.
- 30 AI monetisation opportunities
- Recommended tools for every idea
- Practical playbooks showing how each model works
Join the twice-weekly AI briefing. Free forever, unsubscribe anytime.
Podcast Show Notes Service
Playbook · recommended tools · earning model
AI Affiliate Sites
Playbook · recommended tools · earning model
Children’s Book Publishing
Playbook · recommended tools · earning model
+27 more insideUnlocked instantly when you subscribe