Skip to content

DeepSeek

Route mapped models to DeepSeek's Anthropic-compatible endpoint with a DEEPSEEK_API_KEY.

Updated View as Markdown

DeepSeek serves its models over an Anthropic-compatible endpoint — shunt forwards Claude Code’s Messages request as-is and injects the DeepSeek API key. There is no built-in preset, so the upstream declares kind and base_url explicitly.

Quick start

Let a coding agent wire it up for you — for a provider without a named blueprint, shunt add injects the documentation URL into its generic research guide (offline and read-only; the agent edits the config, the command never does):

shunt add upstream https://api-docs.deepseek.com/guides/anthropic_api --print | claude

Or follow the manual steps below.

Configure the upstream

[[upstreams]]
name = "anthropic"
provider = "anthropic"   # keep Anthropic as the default for unrouted models (e.g. claude-*)

[[upstreams]]
name = "deepseek"
kind = "anthropic"
base_url = "https://api.deepseek.com/anthropic"
auth = { mode = "api_key", env = "DEEPSEEK_API_KEY" }

[[routes]]
model = "deepseek-v4-pro"
provider = "deepseek"

[[routes]]
model = "deepseek-v4-flash"
provider = "deepseek"

Ordered [[upstreams]] replace shunt’s built-in providers, so the config must declare the anthropic default it still falls back to (server.default_provider defaults to anthropic).

The legacy [providers.deepseek] table form remains supported — but do not mix [[upstreams]] and [providers.*] in one file.

Credentials

export DEEPSEEK_API_KEY='...'

Never write the key into the config. shunt check validates the config’s structure but does not read the key’s value — if DEEPSEEK_API_KEY is unset, the first request routed to deepseek returns an authentication error.

Models

Model id Notes
deepseek-v4-pro frontier tier
deepseek-v4-flash fast, lighter tier

Select a routed id in Claude Code via ANTHROPIC_MODEL, ANTHROPIC_CUSTOM_MODEL_OPTION, or a subagent’s model: frontmatter. To surface an entry in the /model picker instead, advertise a claude-prefixed alias with a [models.upstream_model] map — see Model Discovery.

Verify

shunt check    # -> config ok
shunt run
curl -sS http://127.0.0.1:3001/v1/messages \
  -H 'anthropic-version: 2023-06-01' \
  -H 'content-type: application/json' \
  -d '{"model":"deepseek-v4-flash","max_tokens":16,"messages":[{"role":"user","content":"Reply with OK."}]}'

Confirm the response’s x-gateway-upstream header names deepseek, then point Claude Code at shunt.

Subagent plugin

The shunt-deepseek plugin ships one ready-made Claude Code subagent per model above:

/plugin marketplace add pleaseai/shunt
/plugin install shunt-deepseek@shunt
Navigation

Type to search…

↑↓ navigate↵ selectEsc close