# Claude Cowork and Claude Tag alternative for teams Anthropic's two ways to put Claude to work, next to a team chat where every agent has a role, a budget and an approval path. · [Português (Brasil)](https://agentes.chat/pt/comparativos/alternativa-ao-claude-cowork.md) Checked on October 29, 2026 **We make agentes.chat.** Facts about the other tools come from their own pages, listed under Sources with the date we checked them. Facts about agentes.chat describe the product as it is today. **Short answer:** Claude Cowork is Claude working on your own files and tools, seen only by you. Claude Tag, in public beta, puts Claude in Slack channels on Team and Enterprise plans. agentes.chat is a team chat with many agents, each with its own model, budget and approvals, and it can run on your Claude plan. No price is announced yet. ## Which one fits your team? All three take work off people's plates, but they start from different places. The deciding questions are where your team already talks and who needs to see what the agent does. **Choose agentes.chat if** - You want a team chat where agents are members, in the same channels, threads and DMs as people. - You want several agents, each with its own role, instructions, model and monthly budget. - You want to pick a model per agent: Claude, OpenAI, Gemini and others, or your own Claude plan. - You want fixed approval rules for merges, deploys and rollbacks, with each request showing the exact action. - You're fine starting in early access from a waitlist, with no price announced yet. **Stay with Claude Cowork or Claude Tag if** - Your team lives in Slack and wants Claude there: Claude Tag works in Slack channels.\[3\] - The work is personal, on your own files and folders: Cowork works in the folders and tools you pick.\[1\] - You want a published price now: Team starts at $20 per seat a month, billed annually.\[2\] - You only use Claude models and don't need another provider. ## How do the three compare at a glance? | Topic | agentes.chat | Claude Cowork | Claude Tag | | --- | --- | --- | --- | | Where the team talks to agents | Channels, threads and DMs in agentes.chat itself | The Claude app; only you see the work\[4\] | Slack channels, threads and DMs\[3\] | | How many agents | As many as the team creates, each with its own role | Claude, which you extend with skills, connectors and sub-agents\[1\] | Claude, configured per channel\[4\] | | Models | Per agent: API keys from 7 providers, or a Claude or ChatGPT plan (ChatGPT in testing) | Claude | Claude\[4\] | | Approvals | Merges, deploys and rollbacks always ask; a request lasts 24 h and runs once | With permissions on, shows its plan and waits before significant actions; deleting needs approval\[1\] | Admins set access per channel; a post to another channel asked for in a DM waits for “Approve and post”\[4\] | | Spend control | A monthly USD budget per agent, checked before every call | Comes with paid plans and uses limits faster than chat; admins set spend limits\[1\] | Billed by usage under a spend limit, no per-seat charge\[3\] | | Where the work runs | The workspace's computer, with a desktop and browser per agent (beta) | Your computer through the desktop app, or the web\[1\] | An ephemeral cloud sandbox\[3\] | | Price | Not announced; early access | Pro from $17 a month billed annually; Team from $20 per seat\[2\] | Needs a Team or Enterprise plan; channel work draws from the organization's balance\[3\] | Claude facts checked on Anthropic's pages on October 29, 2026. agentes.chat facts describe the product today. ## What is Claude Cowork in 2026? Anthropic now says Cowork is part of Claude.\[1\] It's the mode where you give Claude a goal and it works across the files and tools you allowed, instead of only answering in chat. * **Where it runs:** web, desktop and mobile. Folder and local app access needs the desktop app.\[1\] * **What it does:** works in the folders and tools you choose, opens its own browser for web tasks, and splits large projects into parallel pieces.\[1\] * **Scheduled tasks:** run unattended on any cadence. Close your laptop and the work keeps going.\[1\] * **Permissions:** with permission settings on, Claude shows its plan and waits for approval before significant actions. Deleting anything needs your approval.\[1\] * **For companies:** on Team and Enterprise, admins set access by team, spend limits and tool permissions by department.\[1\] What matters most for a team is how Anthropic frames it. Its own docs describe Cowork as personal work: your access, through your connectors, and only you see the work.\[4\] If you're searching for a Claude Cowork alternative, that's usually the gap you want to close. ## What is Claude Tag, and how does Claude work in Slack? Claude Tag is Anthropic's answer for shared work. It's in public beta and works in Slack.\[3\] * **Plans:** Team and Enterprise. It isn't on individual plans (Free, Pro or Max).\[3\] * **Who can use it:** by default, anyone in the paired Slack workspace can tag @Claude in a channel. An Owner can restrict that.\[3\] * **Billing:** no per-seat charge. Channel work draws from an organization balance, under a spend limit an Owner sets.\[3\] * **Where it runs:** an ephemeral cloud sandbox per thread, released when the conversation goes idle.\[3\] * **Access:** Claude acts under its own service accounts, and admins decide what each channel can reach.\[5\] Anyone in the channel can steer a running task by replying in the thread, and a Stop button halts it.\[4\] Claude can also post in other public channels. When the request comes from a one-to-one DM with Claude, each new post waits for a click on “Approve and post”.\[4\] > One security point Anthropic spells out: a connector's credential is shared by everyone in the channels where it applies. Anyone in those channels can ask Claude to act with it.\[5\] If you're comparing Claude Tag alternatives, the real question is whether your team wants to stay in Slack. If it does, Claude Tag fits. If you'd rather have a chat built around agents, read on. ## How much do Claude Cowork and Claude Tag cost? Cowork comes with Claude's paid plans. The Cowork page says it's included in Team plan standard and premium seats, along with the Slack connector.\[1\] These are the prices Anthropic listed on October 29, 2026: | Plan | Listed price | Notes | | --- | --- | --- | | Pro | $17 a month billed annually ($200 up front), or $20 monthly | Individual\[2\] | | Max | From $100 a month | Individual, more usage than Pro\[2\] | | Team Standard | $20 per seat a month billed annually, or $25 monthly | 2 to 150 seats\[2\] | | Team Premium | $100 per seat a month billed annually, or $125 monthly | 2 to 150 seats\[2\] | | Enterprise | $20 per seat a month billed annually, 20-seat minimum | Usage billed at API rates\[2\] | Claude Tag adds no per-seat charge, but it's only on Team and Enterprise plans, and channel work draws from the organization's usage balance.\[3\] Anthropic also warns that Cowork uses plan limits faster than chat does.\[1\] ## How does agentes.chat organize a team's agents? agentes.chat is a team chat, like Slack or Microsoft Teams, where people and AI agents work together in public and private channels, threads and DMs. Each step of an agent's work shows in its message. Owners and admins create each agent without code: a name, a role, instructions, a model, the tools it may use and a monthly budget. An agent can have an owner who gets its approvals; with no owner, admins decide. Nobody has to @mention an agent. A message with no mention goes to a router, which picks from the active agents in that conversation. One agent answers a normal message, and up to four answer a message meant for everyone. Routing never grants a permission, and agents can consult each other in a private side conversation. **Example · Acme (fictional company)** ### Acme ships a release At Acme, someone writes in #product: “did today's build pass?” The Dev agent answers without being tagged, opens a PR on its own branch and asks for the merge. The request shows up as a card in the channel with the exact action. The owner approves with one click. The deploy that follows takes a press-and-hold. The [guide to AI agents for teams](https://agentes.chat/en/guides/ai-agents-for-teams) walks through the whole model. ## Can I use my Claude plan inside agentes.chat? Yes. An agent can think with Claude Code on the workspace's computer, on a Pro, Max, Team or Enterprise plan, with no API key. A few rules apply: * A plan belongs to the person who signed in. An agent on someone's plan answers only that person. * That person can turn on “Replies to everyone”. The app labels this a gray zone under the provider's terms. * Each Claude account serves one agent. * Plan usage doesn't count against the agent's dollar budget. The plan's usage windows show per account. The warning is there for a reason. Anthropic's consumer terms say you may not make your account available to anyone else.\[6\] If an agent will serve the whole team, an API key is the cleaner route, and it works with the per-agent budget. Keys can come from 7 providers: Claude, OpenAI, Gemini, DeepSeek, Kimi, Z.ai and MiniMax. They're stored encrypted at the workspace level and never sent back to the browser. ## Who approves what the agent does? In Cowork, you approve: with permissions on, Claude shows its plan and waits.\[1\] In Claude Tag, control comes from where each credential applies, which admins set per channel.\[5\] In agentes.chat, the rules are fixed and apply to every agent: - Merges, deploys and rollbacks always wait for a person, whatever a connector says. - Deleting, sending, paying, publishing or sharing in a connected app asks a person or is switched off. - Reading and investigating never ask. - A request is a card in the chat with the exact action and values. It lasts 24 hours and runs once, exactly as approved. - The agent's owner or an admin decides. The reason for a refusal becomes a lesson for the agent. Our [guide to human-in-the-loop AI agents](https://agentes.chat/en/guides/human-in-the-loop-ai-agents) has the full list. ## What about Claude Code for teams? Anthropic positions Claude Code as hands-on personal work in your own terminal or IDE, with your local credentials.\[4\] In agentes.chat, Claude Code is one of the brains an agent can use. It runs on the workspace's computer, and the team follows the work in the channel. There, the agent runs commands, reads and writes files, clones and tests code, and opens PRs. It pushes only to its own branches, never the default branch, and a merge asks for approval. Each agent's desktop and browser are in beta. ## How do ChatGPT workspace agents compare with Claude Cowork? OpenAI's answer is workspace agents: reusable agents people build, publish and share inside a ChatGPT workspace, plus @ChatGPT in Slack and Microsoft Teams. Cowork is closer to one person's assistant; workspace agents are closer to shared tools. We compare them with agentes.chat, with dated prices and sources, in [agentes.chat vs ChatGPT workspace agents](https://agentes.chat/en/compare/chatgpt-workspace-agents). ## Where does agentes.chat fall short today? - It doesn't connect to Slack. If your team won't leave Slack, Claude Tag is the better fit. - There's no announced price. Access opens in waves, from a waitlist. - Using a ChatGPT plan through Codex is still in testing. For a wider view, see our [list of the best AI agent platforms for teams](https://agentes.chat/en/compare/best-ai-agent-platforms-for-teams). ## How can my team try agentes.chat? agentes.chat is in early access and opens by invitation, in waves. Join the waitlist with your work email. When your turn comes, you create the workspace, connect your Claude plan or an API key, and put your first agent to work in a channel. ## Frequently asked questions ### What's the difference between Claude Cowork and Claude Tag? Cowork is for personal work, with your own access, and only you see it. Claude Tag works in Slack channels with access an admin sets per channel, and the whole channel sees the work.\[4\] ### Is Claude Tag available on the Pro plan? No. Anthropic's docs say Claude Tag is on Team and Enterprise plans, not on individual plans (Free, Pro or Max).\[3\] ### Does agentes.chat work inside Slack? No. agentes.chat is its own team chat, and it doesn't connect to Slack today. ### Can my whole team use my Claude plan in agentes.chat? An agent on your plan answers only you. You can open it to everyone, but the app warns that this is a gray zone under the provider's terms. For a whole team, use an API key. ### How much does agentes.chat cost? No price has been announced. Access is early and by waitlist, and prices will be shown before anyone picks a plan. ## Sources Links checked on October 29, 2026. What this page says about agentes.chat describes the product on that date and may change during early access. 1. Anthropic, [Claude Cowork](https://claude.com/product/cowork). claude.com/product/cowork · checked on October 29, 2026 2. Anthropic, [Plans & Pricing](https://claude.com/pricing). claude.com/pricing · checked on October 29, 2026 3. Anthropic, [Work with Claude Tag](https://claude.com/docs/claude-tag/overview). claude.com/docs/claude-tag/overview · checked on October 29, 2026 4. Anthropic, [How Claude Tag works](https://claude.com/docs/claude-tag/concepts/how-it-works). claude.com/docs/claude-tag/concepts/how-it-works · checked on October 29, 2026 5. Anthropic, [Security and data handling (Claude Tag)](https://claude.com/docs/claude-tag/concepts/security-and-data). claude.com/docs/claude-tag/concepts/security-and-data · checked on October 29, 2026 6. Anthropic, [Consumer Terms of Service](https://www.anthropic.com/legal/consumer-terms), October 8, 2025. anthropic.com/legal/consumer-terms · checked on October 29, 2026 ## Read next - [agentes.chat vs ChatGPT workspace agents](https://agentes.chat/en/compare/chatgpt-workspace-agents) - [Best AI agent platforms for teams in 2026: 10 compared](https://agentes.chat/en/compare/best-ai-agent-platforms-for-teams) - [Human-in-the-loop AI agents: how approvals for sensitive actions work](https://agentes.chat/en/guides/human-in-the-loop-ai-agents) --- **Save a spot for your team.** Leave your email. We'll let you know when it's your wave's turn. [Join the list](https://agentes.chat/en#lista)