# Ray vs Claude Tag

> Claude Tag needs Claude seats. Ray needs a Slack channel. Their facts from their own pages, checked August 2026, including where they are the better pick.

Source: https://heyray.io/compare/claude-tag

## What Claude Tag is

Anthropic's own agent inside Slack, which replaced their earlier Claude in Slack app in June 2026 and is in beta for Team and Enterprise customers. You mention it in a channel and it breaks the request into stages, works through them and posts back. It also reads channels passively and decides for itself when to reply, runs routines such as weekday standups and Friday digests, writes to Linear, Asana, Jira and GitHub, and keeps channel and workspace memory.

Source: https://claude.ai

## One message, two outcomes

In #product, Marco says: "Can we get the release notes drafted and the follow-up tickets filed before Thursday?"

**Ray**

1. Picks it up in the channel, on any Slack workspace
2. Shows you the approval card before it writes anything
3. Costs the same whether it ran once or fifty times

**Claude Tag**

1. Does the same work, on Anthropic's strongest model
2. Needs everyone involved to hold a Claude seat
3. Draws down tokens against the org's balance as it works

## Side by side

| | Ray | Claude Tag |
| --- | --- | --- |
| What you buy first | Nothing. Add it to a Slack channel | Claude Team or Enterprise seats for the people who use it |
| Pricing model | Flat by team size, never metered | Per seat, plus token consumption against an org balance |
| Before a write happens | An approval card for that write, every time | The credentials it acts with are granted up front |
| Setting it up | Self-serve, by anyone who can add a Slack app | Admin-created service accounts |
| Which Slack it works on | Any Slack workspace | Workspaces whose company buys Claude |
| Reading channels | Reads what it is invited to and decides for itself | Reads channels passively and decides for itself |
| Writing to your tools | A curated set of tools, each write approval-gated | Linear, Asana, Jira and GitHub among others |
| Maturity | Live, small, and shipping | Beta, from a very large company |

## What Ray costs

Free covers 25 tasks a month plus a one-time 50-task starter grant that never expires. Pro is one flat price for your team's size: $49 a month for 1 to 3 people, $99 a month for 4 to 10 people, $199 a month for 11 to 25 people. No seats, and using Ray more never costs more.

## What Claude Tag costs

There is no rate card for Claude Tag itself. It is in beta, and Anthropic has published neither a per-seat nor a per-token price for it. What is published is what it sits on: Claude Team seats at twenty-five dollars per user a month billed monthly, or twenty billed annually, and the work itself consumes tokens charged against your organisation's balance as it reads channels, holds context and runs tasks. Admins set spend limits at the organisation and channel level, which tells you plainly what kind of bill it is.

Source: https://claude.ai/pricing

## Choose Claude Tag if

- Your company already pays for Claude Team or Enterprise. Claude Tag comes with the stack you have, runs on Anthropic's strongest model, and adding Ray means adding a vendor. That is the strongest argument on this page and it is theirs.
- You want the frontier model doing the reasoning. Ray routes to what fits the task and is honest that it is not always the largest model available.
- You want one vendor for your AI. Buying the assistant, the agent and the model from the same company is a real simplification, and Ray cannot offer it.
- Your admins prefer central control of who gets an agent and what it may touch. Service accounts and org-level spend limits are that, and they are a reasonable thing to want.

## Choose Ray if

- You do not want to buy Claude seats to get an agent in Slack. Ray works on any workspace, and the people it helps do not each need a licence.
- You want the bill to be one number you can predict. A per-seat price with metered consumption on top is two variables, and the second has no published rate.
- You want each write to wait for a yes, rather than handing an agent credentials up front and trusting the configuration.
- You want to try it this afternoon without an admin creating a service account first.

## What Claude Tag does not do

Each checked against their own documentation, August 2026.

- Claude Tag is in beta and limited to Team and Enterprise customers, so it is not something a small team on a free or Pro Slack can simply add today. (https://claude.ai)
- Anthropic publishes no price for Claude Tag itself. The seats are priced and the consumption is billed to an org balance with admin spend limits, but what a month of real use costs is not something we could read at source, so we do not state one. (https://claude.ai/pricing)

## Questions people ask

**Is Ray better than Claude Tag?**

Not at the things a large company builds well, and we will not claim otherwise. Claude Tag runs Anthropic's strongest model and does most of what Ray does. Ray is better when you do not want to buy seats for everyone, want a bill that does not move, and want each write to wait for a yes.

**We already have Claude Enterprise. Should we use Claude Tag?**

Probably, and that is the honest answer. It is in the stack you already pay for. The reasons to look at Ray anyway are the bill's shape, the per-write approval, and that Ray does not need everyone to hold a seat.

**Does Ray use Claude?**

Ray routes to the model that fits the task, and that includes Anthropic's. The difference between the two products is not whose model is underneath, it is what sits around it: the approval step, the flat price and the setup.

**What does Claude Tag actually cost?**

We cannot tell you, because Anthropic has not published it. You buy Claude Team or Enterprise seats, and the agent's work consumes tokens billed to your organisation's balance. Our cost table prices only the seats, which makes their column a floor rather than an estimate.

## Our rule for this page

Every Claude Tag fact here comes from their own published pages, checked August 2026, and we say where they are the better choice. If anything here is wrong, tell us at support@heyray.io and we will fix it.