DeepSeek

DeepSeek's chat and reasoning models reach Memtro through DeepSeek's API, which speaks the OpenAI chat format. Add the key once and DeepSeek answers everywhere Memtro calls a model: chat, agent mode, routing ladders, memory extraction and jobs. Read your data at DeepSeek before you do.

Add the key

Create the key on DeepSeek's platform at platform.deepseek.com/api_keys. DeepSeek's pricing page says usage is deducted from a balance you top up or have been granted.

On Dashboard → Providers pick DeepSeek, 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, and nothing checks that the key is DeepSeek's: an OpenAI key pasted here is saved like any other. Memtro stores the key encrypted, never shows it again, and asks DeepSeek which models it can reach, so a good key comes back as Added deepseek 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 DeepSeek refuses: Saved. Listing models failed: deepseek 401. Check the key, then use Refresh models. Refresh models beside the key asks again.

Memtro calls DeepSeek at https://api.deepseek.com/v1. Leave the optional base URL empty: on hosted Memtro an override is refused, even one naming DeepSeek's own host, with Model endpoint host "api.deepseek.com" is not allowlisted.

Model ids

Id	What it is
deepseek/<model>	One model through your key. Bare ids starting deepseek- are read the same way.
auto/deepseek	Routed by difficulty within DeepSeek. See model routing.

DeepSeek's model names change often, so this page names none: the model list beside the key shows what your key can reach. Memtro's default DeepSeek ladder names two models, a chat model and a reasoning model, chosen when it was written. If they are not in your list, choose your own models per level on the Routing page before relying on auto/deepseek. A named model is used only where an administrator has opened model choice there; otherwise it is answered as memtro (see cost control).

What is different about DeepSeek in Memtro

• No reasoning setting. Memtro passes no thinking or effort setting to DeepSeek: a reasoning_effort on the request, and agent mode's request for deeper thinking, have no effect here. Choose a reasoning model by its id instead.
• No embeddings. A DeepSeek 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 DeepSeek reports. Allowances are set in tokens, and Memtro does not price them: DeepSeek's platform is where the balance is.

Your data at DeepSeek

As of November 2026, DeepSeek's privacy policy says the personal data it collects is processed and stored in the People's Republic of China and kept for as long as you have an account, and it lists training and improving its technology among the purposes. It does not say whether API requests are treated differently, and we found no API-specific statement on training or retention. A request through Memtro usually carries your organisation's memory with it, so settle those questions with DeepSeek before a shared key sends it there. See privacy for what erasure in Memtro cannot reach.

What this is not

This is the API, billed by DeepSeek against the balance on the account that owns the key.