Claude Console (2026): What It Is, API Keys, the Playground and Common Fixes

Most people searching for the Console want one of three things: a key, somewhere to try a prompt, or a fix for a button that won't do anything. This page covers all three in plain words, from Anthropic's own docs and Help Center as of Nov 2026.

The Claude Console in plain English

Anthropic's docs say API keys for the Claude API, also called Anthropic API keys, live in the Claude Console. So if a developer asks for access to the Anthropic Console, they mean the same site. Here's what sits inside it:

• API keys, at platform.claude.com/settings/keys, where you create, disable and delete keys.
• The playground, at platform.claude.com/playground, for trying a model before anyone writes code.
• Workspaces, which the API docs suggest using to separate environments and control spend by use case. Each workspace has an ID in the ID column of the Workspaces page.
• Rate limits, with a Request tier increase button if you need higher limits or a higher monthly spend cap.
• Billing, where your spend limits show.

The API overview explains that limits come in usage tiers. Your organisation is placed on a tier automatically and can move up over time, and each tier sets a maximum monthly spend plus requests and tokens per minute. For what the API itself does, see my explainer on the Claude API.

Console or chat app: which one do you need?

If you just want to talk to Claude, write with it or upload files, you want the app, not the Console. That's the side I cover in my guide to the Claude desktop app. The Console is the engine room behind products built on Claude. A business owner who never writes code can still end up here, because a freelancer, an agency or an automation tool may ask you for a key that belongs to your organisation.

Your chat plan and API usage are set up differently, which trips a lot of people up. I've covered the difference in a side-by-side look at Claude plans and API billing.

Creating an API key, step by step

Here's the short version from Anthropic's Get your Claude API key page, as of Nov 2026. If you want every detail, I've written a full walkthrough on getting a Claude API key.

1. Go to platform.claude.com and sign in, or create an account if you don't have one.
2. Open the API keys page at platform.claude.com/settings/keys.
3. Click Create key, give the key a name and choose an expiration.
4. Set Linked account to yourself or to a service account, and pick a workspace if you want to scope the key to one.
5. Copy the key straight away and store it somewhere safe, such as a secrets manager.

That last step matters more than it looks. The Console shows the full key, which starts with sk-ant-, only once. If you lose it, you can't view it again: you create a new one. Even the Admin API can't help, because it never returns a key's secret value, only a partially redacted hint.

Personal, service account and workspace keys

When you create a key you choose its type, and the type decides what it can do and when it stops working. This is the part most small teams get wrong.

Key type	Acts as	The catch
Personal key	You, with your roles and permissions	Stops working if you lose access to the organisation. If you're re-invited, old keys aren't restored, so you create new ones.
Service account key	A service account, for CI pipelines, production services or agents	An organisation admin creates the service account. The key stops when that service account is archived.
Workspace key (legacy)	No one: it belongs to the workspace	Keeps working after its creator leaves, until it expires, is disabled or deleted, or the workspace is archived.

Anthropic's advice is clear: use a personal key for your own development, and a service account key for anything shared. The docs also warn that a shared personal key acts as one person and breaks when they leave. Picture the freelancer who connected your website chatbot with their own personal key. The day they're removed from your organisation, the chatbot loses its key.

Expiry dates, disabling and deleting keys

Every key gets an expiration when you create it: a preset of 3 hours, 1 day, 7 days or 30 days, a custom duration, or Never for keys kept in a secrets manager and rotated by hand. If your organisation has a maximum expiration policy, Never isn't offered. You can't change the expiration afterwards.

Anthropic emails the key's creator before it expires: 7 days ahead for keys with a lifetime of at least 14 days, and 1 day ahead for keys with a lifetime of at least 7 days. Shorter keys expire without a warning. After that, requests fail with a 401 authentication error and the key can't be reactivated. On the API keys page, Disable is reversible with Re-enable, while Delete is permanent.

The playground, now Workbench has gone

The Claude Help Center says the playground lets developers try Claude models and API features in the Console, and that it replaces Workbench. It's built on the public Messages API, so the request you build is the same request you'd send in code. To use it:

1. Log in to the Claude Console and go to platform.claude.com/playground.
2. If your organisation uses workspaces, choose the one you want to work in.
3. Type a user message in the prompt area, and add a system prompt if you want to set instructions or context.
4. Click Run, then read Claude's response along with the token counts and usage for that request.
5. Edit the prompt and run it again, or switch models and settings such as temperature and maximum output tokens to see how the answer changes.

You can also add tool definitions, ask for structured outputs, view the raw request and response, load example templates, and click the code toggle to export what you've tested as a snippet. Code examples in Anthropic's docs carry an Open in Playground option too.

Here's the honest catch. The playground doesn't save prompt history or evaluate prompts, and it doesn't store your prompts or conversations on Anthropic's servers. Your draft stays in your browser. Saved prompts, prompt versions, evals and prompt sharing aren't part of it, so if your team relied on Workbench, the Help Center points you to the export function in the legacy version to download your data.

When the Create key button is greyed out

If Create key is disabled, the docs say your role might not allow you to create keys. The fix is to ask an organisation admin to change your role, or to create a service account key for your workload.

A related snag catches developers after the key works. If a key isn't scoped to a workspace, every direct request has to send the workspace ID in a header called anthropic-workspace-id. Scope the key to one workspace when you create it and you skip that. The client SDK reads an environment variable called ANTHROPIC_API_KEY automatically, and direct HTTP requests send the key as an Authorization Bearer header.

What I'd set up first as a business owner

I'd create a personal key with a short expiry while you're testing, and switch to a service account key the moment anything is shared, like a website form or an automation the team depends on. I'd skip workspace keys completely: Anthropic itself calls them legacy. Use workspaces to keep testing apart from live work so you can see where spend goes, and copy anything you want to keep out of the playground, because it won't remember it for you. Anthropic's docs change often, so check their own pages for the latest menus.

If you're setting up AI accounts and keys for a team and want it done safely, that's the kind of AI enablement I help with. I run AI training and workshops for teams, and help businesses with AI enablement and workflow automation, so the busywork runs itself and people get time back. Most of that work is with SaaS marketing teams and small businesses. Here's how we can work together.

Frequently asked questions

Is the Claude Console the same as the Anthropic Console?

Yes. Anthropic's docs say API keys for the Claude API, also called Anthropic API keys, live in the Claude Console at platform.claude.com.

Where do I find my Claude API keys?

On the API keys page in the Claude Console, at platform.claude.com/settings/keys. That's where you create, disable and delete them.

Can I see my Claude API key again after I create it?

No. The Console shows the full key, which starts with sk-ant-, only once at creation. If you lose it, create a new key.

What happened to Workbench in the Claude Console?

Workbench is retired. The playground replaces it for trying Claude models and API features, but it doesn't save prompt history or evaluate prompts.

Why can't I create an API key in the Claude Console?

If the Create key button is disabled, your role might not allow it. Ask an organisation admin to change your role or to create a service account key for you.