Use any model you like.
We turn it into a teammate that does the work.
ChatGPT, Claude and Grok sell intelligence. Copilot and Gemini sell speed inside their own suites. Open agents sell raw power you have to secure yourself. Clone is a different category: governed AI teammates, hired for a role and ready on day one.
Four kinds of AI product. One of them is a teammate.
Every product here is good at what it was built for. They were built for different jobs.
Intelligence
A smarter model, plus an agent that works on your files. You bring the prompts and the follow-through.
Suite productivity
AI inside one vendor's apps. Faster inside those walls, weaker outside them.
Raw power
Open-source agents you host and secure yourself. Powerful, and every risk is yours.
LangChain, CrewAI, n8nAI-employee startups
The right idea in a narrow lane: one chat tool, one small team, or a fixed set of helpers.
Governed AI teammates
Clones are hired for a role, know your company, run tested skills, work every channel, and log every action. On whichever model fits the task.
Seven reasons teams pick Clone
The same seven hold against every product on this page.
Hired for a role, not a prompt
Every clone has a job description, responsibilities, skills and KPIs. Pick one from the marketplace, and your Commander recommends who to hire next. Others hand you a blank chat box.
Grounded context, in layers
Knowledge is layered by org, department, role and clone, and pair memory keeps a separate memory for every clone-and-person pair. Your clone knows which Scott you mean.
Skills that are tested software
SkillForge researches, writes and tests each skill against a live deployed container, repairs its own failures, and opens a pull request in your repo. Elsewhere, “skills” are instruction files.
Orchestration across every system
Any skill can call and chain any other in plain language, across Salesforce, HubSpot, Linear, Google, Microsoft and more. Not just inside one company's suite.
Wherever your people already are
Phone number, SMS, email, Slack, Teams, WhatsApp, a live video avatar, and a seat in the meeting. Clones join and run meetings.
Governance in the architecture
Default-deny on every route, object-level permissions, 100% of clone actions logged, credentials encrypted at rest, and clones pinned to specific skill versions.
Any model, predictable cost
GPT, Claude and Gemini run under the hood, so you're never locked in. One flat plan with credits included, not a seat fee plus metered usage.
See them side by side
Pick any product and compare it with Clone, point by point.
Clone, side by side
Pick a product. We'll tell you where it's strong, too.
Claude gives you intelligence. Clone turns it into governed teammates, and can run on Claude.
Their pitch. Excellent models, and Cowork, a desktop agent for knowledge workers that works on your files, runs scheduled tasks and takes plugins.
Where they're strong
- Among the best reasoning and coding models on the market. We use them too.
- Cowork is a real agent, generally available on desktop, with scheduled tasks and plugins.
- Microsoft's own Copilot Cowork runs on Claude models.
Where they stop
- On Claude Enterprise, seats include no usage: all of it, Cowork included, is billed at API rates on top.
- Cowork activity isn't in Claude's audit logs yet. Enterprise can pull Cowork transcripts through the Compliance API.
- Cowork knows the folders you point it at, not your company's knowledge organized by role and department.
- Claude models only, and no role, phone line or seat in the meeting.
Already using Claude?
Your clones can run on Claude. So even everyday writing, research and code go better through a clone: it already has your company's knowledge, its memory of you, and your sent email to draft from. A chat window starts from zero every time. All of it comes on one Clone plan.
Read the full comparison: Clone vs Claude →Bars show how fully each product covers the capability today. DIY: you'd build it yourself.
Same question. Different teammate.
Generic chat searches the internet. A clone starts from your company: its layered knowledge by org, department and role, and a separate memory for every person it works with.
One shared computer, or a teammate with its own keys
Security has to live in the architecture. Clones get access denied by default, permissions set at the object level, and every action on the record.
- Every bot shares the computer's files, browser sessions and credentials.
- Deleting a bot leaves its files and browser sessions behind.
“Do not use separate Bots as a security boundary.”Grok Bot documentation, SpaceXAI
One bill you can predict
Most AI products charge a seat fee, then meter the agent work on top. A Clone plan is one flat price with usage credits included.
Monthly list prices for 100 users as of the date below, before discounts and taxes. Hatched segments are metered and depend on usage.
The whole picture
Nine capabilities, nine products. Pick a column to see the full comparison.
| Capability | CClone | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Role-based teammates, ready on day one | Hired for a role, with responsibilities, skills and KPIs | Blank chat or agent | Prebuilt agents, not roles | General bots | Blank chat | DIY | Agent gallery | Sales and service agents | One shared coworker | HR, IT, finance roles | Fixed helpers |
| Knowledge by org, department and role, plus memory of each person | Layered knowledge plus pair memory | Folders and projects | Microsoft Graph | Per bot | Per-user memory | DIY | Permission-aware search | Salesforce Data Cloud | Shared team memory | Connected data sources | Per workspace |
| Builds and tests its own skills | SkillForge: tested in a live container | Skill files (instructions) | Copilot Studio (metered) | Skills | Skills and plugins | Self-learning, untested | No-code agent builder | Agent Builder actions | Learns team skills | Workflow builder | Templates and integrations |
| Phone line, SMS, and a seat in your meetings | Phone, SMS, email, chat, video, meetings | Not offered | Not offered | Not offered | Not offered | DIY | Not offered | Salesforce service channels | Slack and Teams | Chat, Slack/Teams, voice | Varies; often one channel |
| Works across systems, not one vendor's suite | Salesforce, HubSpot, Google, Microsoft and more | Connectors | Weaker outside Microsoft | Connectors | Connectors | DIY | Google-centered | Salesforce-first | 3,000+ tools | 250+ integrations | Integrations |
| Separate permissions for each agent | Per clone, default-deny | Per user | Within Microsoft 365 | One shared computer | Per user | Not built in | Workspace permissions | Salesforce permissions | Workspace sandbox, admin scopes | RBAC/ABAC, on-prem option | Varies |
| Every action logged | 100% of clone actions | Cowork not in audit logs yet | Purview | Partial | Compliance API | Not built in | Workspace audit logs | Salesforce audit trail | Enterprise logs; retention on roadmap | Decision traces, config logs | Varies |
| Choice of model | GPT, Claude, Gemini and more | Claude only | Microsoft-curated | Grok only | OpenAI only | Any model | Mostly Gemini | Salesforce-curated | Claude Opus default, vendor-set | Auto-routes many models | Vendor's choice |
| Predictable pricing | Flat plan, credits included | Enterprise: seat plus API-rate usage | License plus credits plus Azure | Seat or bundle | Seat plus credits | Free, plus your engineers | Seat plus usage | Licenses plus per-action credits | Usage credits, from $50/mo | Outcome-based, unpublished | Seat or credits |
Which one fits you?
An honest guide. Sometimes the answer isn't us.
- A sharp assistant for your own writing, research and codeYour clone, running on Claude or GPT
- Drafting faster inside Word, Excel and OutlookMicrosoft 365 Copilot
- Agents for work that never leaves SalesforceSalesforce Agentforce
- An open agent you'll run and secure on your own machineOpenClaw or Hermes
- A lightweight assistant for a small team's inbox and calendarAn app like Lindy
- Teammates hired for real roles, doing the work across your systems and channels, with governance built inClone
Could you build it yourself?
With LangChain, CrewAI, n8n or OpenClaw, you can build an agent. To match a clone, you'd also need to build and run:
- Layered knowledge for each org, department and role
- Memory for every clone-and-person pair
- A skill pipeline that tests against live systems and repairs itself
- Object-level permissions, default-deny on every route
- An audit log of every action
- Phone numbers, SMS and a meeting bot
- Routing across model providers
- Credential vaulting, version pinning and failover
Or turn on a team that already has all of it.
Questions buyers ask
Is Clone another AI model?
No. Clone runs on the leading models (GPT, Claude and Gemini) and uses the one that fits each task. What Clone adds is the teammate around the model: a role, your company's knowledge, skills it actually runs, the channels your people use, and governance on every action.
We already use ChatGPT or Claude. Why move to Clone?
Because your clones run on those same models and add everything a chat window lacks. Even for everyday writing, research and code, a clone starts from your company's knowledge, its memory of you and your sent email, where chat starts from a blank box. Clones also own outcomes: they work your inbox, phone line, CRM and meetings, and log every action for review. All of it on one Clone plan with usage included.
How is Clone different from Microsoft 365 Copilot?
Copilot makes people faster inside Word, Excel, Outlook and Teams. Clones do the work itself, inside Teams and Outlook and across every other system, including Salesforce, HubSpot and Google. They take calls, join meetings, and come on one flat plan with usage included, instead of a license plus metered agent charges.
How is Clone different from Grok Bot?
Grok's documentation says every bot on an account shares one cloud computer, including its files, browser sessions and credentials, and that separate bots shouldn't be used as a security boundary. Each clone has its own permissions, default-deny access and its own audit trail.
Couldn't we build this ourselves with OpenClaw, LangChain or n8n?
You could build an agent. Matching a clone means also building layered company knowledge, per-person memory, a skill pipeline that tests against live systems, object-level permissions, full audit, phone and meeting channels, and model routing, then securing and running all of it. Clone gives you the power of an open agent with those guardrails already built.
How does Clone pricing compare?
Clone plans are a flat monthly price with usage credits included. For example, Pro+ is $2,000 a month for up to 100 users, with $2,000 in credits included. Most alternatives charge a seat fee and then meter agent usage on top.
Which models does Clone use?
GPT, Claude and Gemini all run under the hood, so you're never locked into one provider.
How does Clone handle security?
Security is built into the architecture: access is denied by default on every route, permissions are set at the object level across orgs, departments, clones, skills and users, 100% of clone actions are logged, credentials are encrypted at rest, clones are pinned to specific skill versions, and each customer's data is kept separate. Clone runs on Google Cloud with Azure failover. Our SOC 2 Type II audit is in progress.
Bring the models you love.
Get the team that does the work.
Clones run on GPT, Claude and Gemini, in the tools you use, on one plan.