the provider catalog

OpenAI

{"meta":{"title":"OpenAI · typed outputs in Nika workflows","angle":"The most widely known API. Structured outputs make schema: a real contract. The answer arrives as data the next task can trust.","pick":"when downstream tasks consume the answer as data, not prose.","console":{"label":"platform.openai.com","href":"https://platform.openai.com/api-keys"},"setup":[{"text":"Create a key in the platform console."},{"text":"The key rides the environment, never the file","code":"export OPENAI_API_KEY=…"}]},"usage":{"id":"openai","yaml":"nika: v1\nworkflow:\n  id: typed-extraction\n  description: \"the answer arrives as data, not prose — schema makes the shape a contract\"\n\ntasks:\n  extract:\n    infer:\n      model: openai/gpt-5-mini\n      max_tokens: 200\n      schema:\n        type: object\n        required: [severity, component]\n        properties:\n          severity: { type: string, enum: [low, medium, high] }\n          component: { type: string }\n      prompt: |\n        \"Login fails with a 500 after the session-store deploy.\"\n        Extract severity and component.\n\noutputs:\n  ticket: ${{ tasks.extract.output }}\n","file":"openai.nika.yaml"},"audit":{"clean":true,"models_resolve":true,"tasks":1,"waves":1,"cost_ceiling_usd":0.0004,"cost_floor_usd":0.0004,"pricing":[{"model":"openai/gpt-5-mini","input_per_million":0.25,"output_per_million":2}],"pricing_as_of":"2026-07-07"}}

[{"slug":"gpt-5-mini","id":"gpt-5-mini","aliases":["gpt-5-mini"]},{"slug":"gpt-5-2","id":"gpt-5.2","aliases":["gpt-5.2"]}]

The most widely known API. Structured outputs make schema: a real contract. The answer arrives as data the next task can trust. 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.

2pinned modelsaliases pin exact upstream ids
cloudfamilybring your own key; it rides an env var and stays yours
OPENAI_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: openai/gpt-5.2), and the engine speaks openai-chat on the wire. The key rides OPENAI_API_KEY in the environment, never a config file, never the workflow. The provider id also answers to gpt.

default: gpt-5.2
cheap: gpt-5-mini
dialect: openai-chat
tags: code-execution · frontier · function-calling · image-gen · prompt-caching · realtime · streaming · vision

the models

2 pins

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.

gpt-5.2gpt-5.2ctx 272k · reasoning · vision · json:schema
gpt-5-minigpt-5-minictx 272k · reasoning · vision · json:schema$0.25/M in · $2/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

2 rooms

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.

gpt-5-minias gpt-5-mini
gpt-5.2as gpt-5.2

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

in a real file

a complete workflow · shown whole

openai.nika.yaml

1nika: v12workflow:3  id: typed-extraction4  description: "the answer arrives as data, not prose — schema makes the shape a contract"56tasks:7  extract:8    ◇infer:9      model: openai/gpt-5-mini10      max_tokens: 20011      schema:12        type: object13        required: [severity, component]14        properties:15          severity: { type: string, enum: [low, medium, high] }16          component: { type: string }17      prompt: |18        "Login fails with a 500 after the session-store deploy."19        Extract severity and component.2021outputs:22  ticket: ${{ tasks.extract.output }}23

check-green against the released engine · 1 task · 1 wave · worst-case ceiling $0.0004 · the drift gate re-proves this copy on every test run · open it in the playground →

get set up

keys at platform.openai.com

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

Pick it for when downstream tasks consume the answer as data, not prose. Then prove the seat: nika doctor names every provider this binary can reach.

swap it

10 more cloud seats

Moving off OpenAI 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: openai/gpt-5.2 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 →