DeepSeek's official docs keep `/chat/completions` stateless and require full history replay for every multi-round app turn

Checked on June 27, 2026: DeepSeek's official multi-round conversation guide says `/chat/completions` does not store context server-side, so each follow-up call must resend the prior conversation history, including earlier assistant output, before the next user turn.

Accepted official-source monitoring note

Today's accepted item stays DeepSeek-first and uses a current official DeepSeek documentation page because the official @deepseek_ai X timeline still was not safely readable in this run. DeepSeek's English homepage continues to anchor the official X account through the V4 Preview banner, but the direct timeline did not expose a reliably crawlable current post here, so the publish-safe choice is a first-party docs-backed developer update.

What we verified on June 27, 2026

DeepSeek's official multi-round conversation guide says /chat/completions is stateless: the server does not keep prior turn context for you.
The official fix is explicit rather than implied: each later request must concatenate all previous conversation history and send it back to the API.
DeepSeek's own example appends the assistant response from round one before sending round two, then adds the next user question after that.
The published sample still uses the ordinary OpenAI-compatible route with base_url=https://api.deepseek.com and model=deepseek-v4-pro, which means this is not a special SDK-only edge case.
The guide makes the support boundary clearer for production apps: if a follow-up answer looks forgetful, the first debugging target is usually the caller's messages replay, not a hidden DeepSeek memory failure.

Why this is publishable

This is a current official DeepSeek API guide, not a rumor, repost, or inventory claim.

It gives a first-party answer to one of the most common multi-turn support bugs: assuming DeepSeek stores chat state server-side.
It is concrete enough to shape real implementation behavior because the example shows the exact replay order for assistant and user turns.
It stays distinct from the already-published agent-integration news run because the story is about API state handling, not another terminal or desktop wrapper.
It does not imply any new model inventory, pricing tier, or stocked /pricing card.

Why this matters for DeepSeek-first SEO pages

Support pages can now target DeepSeek stateless chat history with official wording instead of generic OpenAI-compatible guesses.
Teams building chat or agent loops can separate conversation replay bugs from thinking-mode or tool-call bugs, which are different failure classes.
DeepSeek-first implementation pages can keep the headline on DeepSeek's API contract rather than drifting into broader non-DeepSeek chatbot theory.

Rejected candidates today

Official X timeline as the primary source: rejected for this run because the homepage still anchored DeepSeek's X presence, but the direct X surface remained unreadable here.
The homepage V4 Preview X anchor: official, but older and already covered by existing V4 launch and migration pages.
The current agent-integration pages: official, but already represented across the site's recent daily news set and therefore higher duplicate-content risk.
The official JSON Output guide: official and useful, but stronger as a long-tail support page than as the single daily news headline.
The official change log and status page: both checked, but neither contained a stronger current developer signal than the multi-round conversation contract.

Editorial takeaway

The safest official DeepSeek story today is a multi-round API contract check: DeepSeek's own docs say /chat/completions remains stateless, so each new turn must replay the full prior history, including assistant output. That is a stronger current developer signal than forcing another already-published integration headline or recycling the older X anchor we still could not fully verify live.