the provider catalog

DeepSeek

{"meta":{"title":"DeepSeek · frontier quality at floor prices","angle":"Strong models at aggressive prices, with an open-weight track record. The audit’s cost ceiling stays almost invisible.","pick":"high-volume text work where the price per call decides.","note":"DeepSeek · Hangzhou, China.","console":{"label":"platform.deepseek.com","href":"https://platform.deepseek.com/api_keys"},"setup":[{"text":"Create a key in the platform console."},{"text":"The key rides the environment, never the file","code":"export DEEPSEEK_API_KEY=…"}]},"usage":{"id":"deepseek","yaml":"nika: v1\nworkflow:\n  id: bounded-answer\n  description: \"a tight token budget makes the audit price a hard ceiling, not a guess\"\n\ntasks:\n  answer:\n    infer:\n      model: deepseek/deepseek-chat\n      max_tokens: 80\n      prompt: |\n        In one sentence: why should a workflow engine refuse\n        to run a file it cannot price?\n\noutputs:\n  answer: ${{ tasks.answer.output }}\n","file":"deepseek.nika.yaml"},"audit":{"clean":true,"models_resolve":true,"tasks":1,"waves":1,"cost_ceiling_usd":0.000022400000000000002,"cost_floor_usd":0.000022400000000000002,"pricing":[{"model":"deepseek/deepseek-chat","input_per_million":0.14,"output_per_million":0.28}],"pricing_as_of":"2026-07-07"}}

[{"slug":"deepseek-chat","id":"deepseek-chat","aliases":["chat"]}]

Strong models at aggressive prices, with an open-weight track record. The audit’s cost ceiling stays almost invisible. One provider/model line moves a workflow here or away again: the same file runs on every provider the catalog names, and swapping is a one-word diff. Machines read the catalog; the binary answers nika catalog.

1pinned modelaliases pin exact upstream ids
cloudfamilybring your own key; it rides an env var and stays yours
DEEPSEEK_API_KEYthe key rides herean env var, never a config file
openai-chatwire dialectthe adapter speaks it

the wire

cloud · bring your own key; it rides an env var and stays yours

A workflow names this provider in the model line (model: deepseek/deepseek-chat), and the engine speaks openai-chat on the wire. The key rides DEEPSEEK_API_KEY in the environment, never a config file, never the workflow. The provider id also answers to deep-seek.

default: deepseek-chat
dialect: openai-chat
tags: budget · chinese · code · function-calling · open-source · reasoning · streaming · structured-output

the models

1 pin

Aliases are what the file may write; the engine pins each one to an exact upstream id, so a run stays reproducible after a vendor rotates names.

chatdeepseek-chatctx 64k · json:object$0.14/M in · $0.28/M out

prices are the ENGINE's, not ours: the audit's pricing snapshot (models.dev · pinned 2026-07-07): the same table nika check prices your ceiling with.

in the released catalog

1 room

The wire models this provider serves in the released binary's catalog · each one is a room with its seats, exact prices and measured energy.

deepseek-chatas chat

the whole register: /catalog/models · what the released binary knows, vendored at the engine pin.

in a real file

a complete workflow · shown whole

deepseek.nika.yaml

1nika: v12workflow:3  id: bounded-answer4  description: "a tight token budget makes the audit price a hard ceiling, not a guess"56tasks:7  answer:8    ◇infer:9      model: deepseek/deepseek-chat10      max_tokens: 8011      prompt: |12        In one sentence: why should a workflow engine refuse13        to run a file it cannot price?1415outputs:16  answer: ${{ tasks.answer.output }}17

check-green against the released engine · 1 task · 1 wave · worst-case ceiling $0.0000 · the drift gate re-proves this copy on every test run · open it in the playground →

get set up

keys at platform.deepseek.com

1Create a key in the platform console.
2The key rides the environment, never the file export DEEPSEEK_API_KEY=…

Pick it for high-volume text work where the price per call decides. DeepSeek · Hangzhou, China. Then prove the seat: nika doctor names every provider this binary can reach.

swap it

10 more cloud seats

Moving off DeepSeek is editing one line. The rest of the file (tasks, permits, outputs, the trace) does not change. The register orders the catalog local first; beyond it the engine embeds 21 more OpenAI-compatible endpoints.

model: deepseek/deepseek-chat model: ollama/qwen3.5:4bthe whole migration · nothing else moves

cross-references

the words that meet it

[]

where it lives

Try it in the playground, or install and go local first. Read the spec →