Anthropic (Claude)

Anthropic's Claude models reach Memtro through the Messages API with a Claude Console API key. Add the key once and Claude answers everywhere Memtro calls a model: chat, agent mode, routing ladders, memory extraction and jobs.

Add the key

Create the key in the Claude Console at platform.claude.com/settings/keys: choose Create key, name it, pick an expiry and the account it is linked to (you, or a service account), and optionally a workspace. The key starts with sk-ant- and is shown once. A key linked to you stops working if you leave the Anthropic organisation, so a service account suits a key the whole organisation will share. Anthropic's getting started page has the details.

On Dashboard → Providers pick Anthropic, paste the key and choose who it is for: Personal (only you), Shared with your organisation, or one team. Which of these you are offered depends on your permissions. Work done for the organisation or a team, such as its jobs and its connections' syncs, spends only shared and team keys, never a member's personal one; your own personal jobs and personal connections can use your personal key.

Saving does not send a test request. Memtro stores the key encrypted, never shows it again, and asks Anthropic which models it can reach, so a good key comes back as Added anthropic key (personal). with the number of models it found after it. If the listing fails, the key is still saved and the form says why, for a key Anthropic refuses: Saved. Listing models failed: anthropic 401. Check the key, then use Refresh models. Refresh models beside the key asks again.

A Claude Code token (sk-ant-oat…) is refused here with That is a Claude Code token, not an API key. Tokens from a Claude subscription belong under the Claude Code provider: see the Claude Code guide.

Leave the optional base URL empty. Hosted Memtro accepts an override only when it is https:// on one of the three hosts Memtro speaks to, api.anthropic.com, api.openai.com or generativelanguage.googleapis.com, so an override cannot point this key at a proxy or gateway; any other host is refused. A host your organisation approves under Custom model hosts does not count here: that list is for OpenAI-compatible endpoints only.

Model ids

Id	What it is
anthropic/<model>	One model through your key, e.g. anthropic/claude-sonnet-5 or anthropic/claude-opus-5. A bare id that starts with claude is read the same way.
auto/anthropic	Routed by difficulty within Anthropic, on the default ladder Haiku 4.5 → Sonnet 5 → Opus 5 → Fable 5.1. See model routing.

The model list beside the key shows every id the key can reach. A named model is used only where an administrator has opened model choice on the Routing page; otherwise it is answered as memtro (see cost control).

What is different about Anthropic in Memtro

• Prompt caching. Memtro marks the stable front of each request for Anthropic's prompt cache with a one-hour lifetime: its own instructions, up to two of your system blocks that carry cache_control, and the conversation so far. A follow-up within the hour reads that part from the cache rather than as fresh input, and the cached reads Anthropic reports are recorded with the request.
• Tool search. Anthropic is the one provider that receives Memtro's connector tools for native deferred loading, so the model loads a tool's definition when it needs it. Every other provider gets one combined tool per connected system.
• Thinking. Agent mode, and a reasoning_effort of high, xhigh or max (deep) or low or medium (standard), become Anthropic's extended thinking with an effort level. A request capped below 2,048 output tokens gets no thinking.
• Token counts. POST /v1/messages/count_tokens uses Anthropic's own tokeniser for a request bound for an Anthropic model. See the API reference.
• No embeddings. An Anthropic key does not embed memories. Memory uses an OpenAI or Google key your organisation holds, or Memtro's own embedding model.
• More than one key. Your own key is tried first, then your team's, then the organisation's; within each, the order set on the Providers page, or Spread the load. A rate-limited or exhausted key hands over to the next before routing moves down the ladder. See several keys for one provider.
• Usage in tokens. Each request is recorded with the input and output tokens Anthropic reports. Allowances are set in tokens, and Memtro does not price them: Anthropic's own console is where the bill is.

Your data at Anthropic

As of November 2026, Anthropic says that by default it will not use inputs or outputs from its commercial products to train models, and that API inputs and outputs are deleted from its backend within 30 days unless you have agreed otherwise, for example zero data retention. Bear in mind that a request through Memtro usually carries your organisation's memory with it. Erasing something in Memtro does not reach what Anthropic already holds: see privacy.

What this is not

This is the API, billed by Anthropic per token on the account that owns the key. It is not a Claude subscription: to run Memtro on one, add a token under the Claude Code provider instead, as the Claude Code guide describes.