OpenAPI spec for OpenAI model, chat completion, and fine-tuning automation via API
OpenAI provides hosted large language models and generative AI capabilities — chat and text completions, embeddings, image generation, speech-to-text and text-to-speech, moderation, fine-tuning, and the Assistants/Responses APIs for building conversational and agentic applications.
This project provides OpenAPI specs for automating against the OpenAI REST API via an Integration Model. The -latest spec is a curated subset covering common CRUD for automation — see OpenAPIs below.
| Asset | Description |
|---|---|
| OpenAPIs/ | OpenAI REST API OpenAPI specs — curated -latest plus the full dated spec |
| Studio Projects/OpenAI | 22 workflows covering common CRUD automation |
| Requirement | Version |
|---|---|
| Itential Platform | 6.x |
| OpenAI API | 2.3.0 |
| OpenAI Integration Model | Required to build automation against the OpenAPI specs |
Import one of the OpenAPI specs from OpenAPIs/ as an Integration Model in Admin > Integrations, then create an integration pointing at the OpenAI API.
Authentication is a bearer token in the Authorization header:
Authorization: Bearer <your-openai-api-key> Generate an API key at platform.openai.com/api-keys.
The instance's authentication/server properties should look like this once configured:
{
"authentication": {
"ApiKeyAuth": "<your-bearer-token>"
},
"server": {
"protocol": "https",
"host": "api.openai.com",
"base_path": "/v1"
}
} | Spec | Version | Operations | Description |
|---|---|---|---|
openai-latest.json | latest (curated) | 84 | Trimmed to 84 of 242 upstream operations — see breakdown below |
openai-2.3.0.json | 2.3.0 | 242 | Full, unmodified vendor spec |
openai-latest.jsonActively-maintained spec (x-vendor-api-version: 2.3.0). Trimmed to 84 of 242 upstream operations covering common CRUD for automation.
Resources included, by category:
Excluded: administration/org-management (API keys, users, roles, groups, invites, projects, audit logs, usage/cost reporting), Realtime voice API, ChatKit, Containers (code interpreter sandboxes), Evals, Skills, Uploads (large multipart upload flow), voice cloning/consents, webhooks, and niche video sub-features (characters, remix, extensions, edits). See the repo README for the full scope, or pull the full spec below if you need one of the excluded areas.
openai-2.3.0.jsonFull, unmodified vendor spec covering all 242 upstream operations — the vendor's complete API surface, preserved as-is. See openai-latest.json above for the curated subset if you just need common CRUD automation.
Backed by the OpenAI:latest Integration Model (see openai-latest.json above). The project contains 22 workflows organized into 5 folders, one atomic workflow per API operation, covering the common-CRUD subset of the curated spec.
| Folder | Workflows | Scope |
|---|---|---|
| Models | List, Retrieve, Delete Model | Model catalog |
| Chat Completions | List, Create, Get, Update, Delete Chat Completion | Chat completion lifecycle |
| Files | List, Create, Retrieve, Delete File, Download File Content | File lifecycle |
| Fine-tuning Jobs | List, Create, Retrieve, Cancel Fine-tuning Job | Fine-tuning job lifecycle |
| Assistants | List, Create, Get, Modify, Delete Assistant | Assistant lifecycle |
| Dependency | Notes |
|---|---|
OpenAI:latest Integration Model | Import from openai-latest.json before importing the project |
OpenAI integration instance | Create in Admin > Integrations with the connection properties above. Workflows are wired to an integration instance named OpenAI — update the adapter_id value in each workflow task if yours is named differently |