Key facts
| Provider | Azure OpenAI — OpenAI models hosted in Azure under your subscription |
| API style | OpenAI-shaped requests with deployment paths, api-version parameters and Entra auth |
| Plugsky API | OpenAI-compatible /v1/chat/completions — change the base URL, keep your SDK |
| Models | 30+ models from free to frontier behind one API key |
| Pricing | Flat monthly plans with unlimited fair-use usage; no per-token billing on self-serve |
| Free tier | Free plan with plugsky-micro and plugsky-lite, no card; 14-day full-access trial |
| Deployment | Plugsky cloud, your VPC, on-prem or air-gapped; region choice for residency |
| Live vs roadmap | Chat, streaming, JSON mode, function calling, embeddings, RAG, agents live; audio, images, moderation, files, batch, fine-tuning, assistants, responses coming soon |
TL;DR
- Both speak OpenAI-style requests; the differences are auth, paths and catalogue.
- Azure adds Entra ID, private networking and regional deployment controls.
- Plugsky adds 30+ models, flat monthly self-serve pricing and OpenAI drop-in calls.
- Free tier uses plugsky-micro and plugsky-lite; a 14-day full-access trial covers paid models.
- Honest trade-off: Azure's compliance and identity integrations remain deeper.
How it works, step by step
- List the Azure OpenAI deployments your applications reference and their regions.
- Record which controls are contractual: Entra ID, private endpoints, regional processing, content filters.
- Create a Plugsky account and map each deployment name to a Plugsky model.
- Swap the client configuration: base URL, API key header and path format.
- Run your evaluations and confirm moderation and refusal behaviour on your prompts.
- Move traffic service by service, keeping the Azure configuration for rollback.
Original data
Try it yourself
Open the Azure OpenAI cost calculator →
How the Azure OpenAI endpoint differs
Azure OpenAI is OpenAI-shaped but not identical. Requests target a resource endpoint, include a deployments/<name> path segment and an api-version parameter, and authenticate with either an API key or Microsoft Entra tokens. Model versions are chosen at deployment time, and content filtering is applied by the service.
For teams already inside Azure, that is convenient plumbing: keys live in Key Vault, networking goes through private endpoints, and access is governed by the same policies as the rest of the estate. For teams that are not, it adds a cloud dependency and deployment ceremony around every model change.
How Plugsky's endpoint differs
Plugsky is deliberately plain. You send OpenAI-format chat completions requests to api.plugsky.com/v1 with an API key. Model choice is a parameter, not a deployment object, so trying a new model is a one-line change.
The catalogue covers 30+ models, from free chat tiers to frontier reasoning, all behind one key. Self-serve plans are flat monthly with unlimited fair-use usage (live pricing), which removes per-token forecasting from the critical path. The free plan includes plugsky-micro and plugsky-lite, and a 14-day full-access trial covers paid models. Chat, streaming, JSON mode, function calling, embeddings, RAG and agents are live; audio, images, moderation, files, batch, fine-tuning, assistants and responses are coming soon.
What actually changes when you migrate
The migration is small if your code is already abstracted behind a client wrapper — and tedious if endpoint details are scattered through the codebase.
- Authentication: Azure keys or Entra tokens become a Plugsky API key.
- Paths: remove the deployment and api-version segments; keep the request body.
- Model mapping: replace deployment names with Plugsky model identifiers.
- Filters: re-test content behaviour, since filtering rules differ by platform.
Where Azure still wins: Entra ID integration, private networking and Microsoft compliance programs. Keep Azure for workloads where those are prerequisites; use Plugsky where portability and catalogue breadth lead.
Honest comparison
| Capability | Plugsky | Azure OpenAI | Building in-house |
|---|---|---|---|
| API style | OpenAI-compatible drop-in (base URL + model name) | OpenAI-shaped with deployment paths and api-version | You define the schema |
| Authentication | API keys managed in the dashboard | API keys or Microsoft Entra tokens | You build identity |
| Model catalogue | 30+ models, one key, model as a parameter | OpenAI models, versions fixed per deployment | You host each model |
| Billing | Flat monthly, unlimited fair use (see live pricing) | Usage-based, billed through Azure | GPU + ops cost |
| Residency | Region choice, VPC, on-prem, air-gapped | Azure regions and Azure-native controls | You control the infrastructure |
| Honest gap | Audio, images, batch, fine-tuning coming soon | Entra ID, private networking and compliance depth | You build it |
Frequently asked questions
Is Azure OpenAI the same API as OpenAI?
It is OpenAI-shaped but not identical: Azure uses resource endpoints, deployment paths, an api-version parameter and Entra or key authentication.
Can I move Azure OpenAI calls to Plugsky easily?
Yes, if your code is behind a client wrapper. Change auth, drop the deployment and api-version path segments, and map deployment names to model IDs.
Does Plugsky support private networking?
Plugsky supports VPC, on-prem and air-gapped enterprise deployments, which covers teams that need private connectivity and residency.
Is there a free plan?
Yes — two free AI models, plugsky-micro and plugsky-lite, with no credit card, plus a 14-day full-access trial for paid models.
How is Plugsky billed?
Self-serve plans are flat monthly with unlimited fair-use usage; there is no per-token billing on self-serve. See the live pricing page.
What about content filtering?
Filtering rules differ between platforms. Re-test moderation-sensitive prompts after migrating rather than assuming identical behaviour.
What does Azure OpenAI still do better?
Deep Microsoft identity, networking and compliance integration, plus region-specific deployment controls that Plugsky does not replicate.