DeepSeek Provider Setup

DeepSeek offers capable AI models through an API that follows OpenAI's conventions.

Property	Value
Provider	`deepseek`
Auth	`DEEPSEEK_API_KEY`
API	OpenAI-compatible
Base URL	`https://api.deepseek.com`

Install plugin

Get the official plugin installed, then restart Gateway:

openclaw plugins install @openclaw/deepseek-provider
openclaw gateway restart

Getting started

Get your API key

Head to platform.deepseek.com to generate an API key.

Run onboarding

openclaw onboard --auth-choice deepseek-api-key

It asks for your API key and makes deepseek/deepseek-v4-pro the default model.

Verify models are available

openclaw models list --provider deepseek

If you want to see the plugin's static catalog without a running Gateway:

openclaw models list --all --provider deepseek

Non-interactive setup

For automated or headless setups, supply every flag directly:

openclaw onboard --non-interactive \
  --mode local \
  --auth-choice deepseek-api-key \
  --deepseek-api-key "$DEEPSEEK_API_KEY" \
  --skip-health \
  --accept-risk

Warning

When Gateway runs as a daemon (launchd/systemd), confirm that DEEPSEEK_API_KEY is reachable by that process (for instance, in ~/.openclaw/.env or through env.shellEnv).

Built-in catalog

Model ref	Name	Input	Context	Max output	Notes
`deepseek/deepseek-v4-flash`	DeepSeek V4 Flash	text	1,000,000	384,000	Fast V4 thinking-capable surface
`deepseek/deepseek-v4-pro`	DeepSeek V4 Pro	text	1,000,000	384,000	Default; strongest V4 model
`deepseek/deepseek-v4-flash-vision-exp`	DeepSeek V4 Flash Vision (Experimental)	text, image	1,000,000	384,000	Experimental image understanding

Warning

As of July 24, 2026 at 15:59 UTC, DeepSeek has discontinued deepseek-chat and deepseek-reasoner. Those model IDs no longer work. Point any configured model refs at deepseek/deepseek-v4-flash or deepseek/deepseek-v4-pro instead.

OpenClaw's local cost figures are only estimates. The vision model's built-in estimate is based on DeepSeek's peak pricing; the published off-peak rates come in at half of those amounts. DeepSeek may adjust pricing at any time, and their Models & Pricing page is the final authority for billing.

For image inputs, pick deepseek/deepseek-v4-flash-vision-exp. The standard Flash and Pro models handle text exclusively. DeepSeek's experimental vision model processes PNG, JPEG, GIF, and WebP images via the same API and API key. Check DeepSeek vision for image limits.

Tip

V4 models honor DeepSeek's thinking setting. OpenClaw also replays DeepSeek reasoning_content on subsequent turns, so thinking sessions that involve tool calls can keep going. Pair /think xhigh or /think max with DeepSeek V4 models to ask for DeepSeek's highest reasoning_effort; both resolve to "max".

Thinking and tools

For DeepSeek V4 thinking sessions, replayed assistant messages from a thinking-enabled turn must carry reasoning_content on follow-up requests. OpenClaw's DeepSeek plugin fills in that field automatically, so regular multi-turn tool use works on deepseek/deepseek-v4-flash, deepseek/deepseek-v4-flash-vision-exp, and deepseek/deepseek-v4-pro even when the history originated from another OpenAI-compatible provider (without native reasoning_content) or from a plain assistant message. No /new is needed after switching providers mid-session.

With thinking turned off (including the UI None option), OpenClaw transmits thinking: { type: "disabled" } and removes replayed reasoning_content from outgoing history, keeping the session on the non-thinking DeepSeek path.

New onboarding picks the more capable deepseek/deepseek-v4-pro model. Reach for deepseek/deepseek-v4-flash when cutting cost or latency matters more than peak capability.

Live testing

To run just the DeepSeek V4 direct-model checks from the modern model live suite:

OPENCLAW_LIVE_PROVIDERS=deepseek \
OPENCLAW_LIVE_MODELS="deepseek/deepseek-v4-flash,deepseek/deepseek-v4-pro" \
pnpm test:live src/agents/models.profiles.live.test.ts

Confirms both V4 models finish and that thinking/tool follow-up turns retain the replay payload DeepSeek expects.

To exercise the experimental vision model with the same DEEPSEEK_API_KEY:

OPENCLAW_LIVE_DEEPSEEK_MODEL=deepseek-v4-flash-vision-exp \
pnpm test:live extensions/deepseek/deepseek.live.test.ts

This covers text, generated-image recognition, and thinking replay checks against the chosen model.

Config example

{
  env: { vars: { DEEPSEEK_API_KEY: "sk-..." } },
  agents: {
    defaults: {
      model: { primary: "deepseek/deepseek-v4-pro" },
    },
  },
}

Model selection, Picking providers, model refs, and failover behavior.
Configuration reference, Complete config reference for agents, models, and providers.